DEV Community: Chime

How B2B founders find the right LinkedIn posts to comment on

Chime — Sun, 14 Jun 2026 03:04:14 +0000

Most B2B founders we work with have the same problem: they know commenting on LinkedIn works, but they spend 30 minutes finding a decent post to comment on and still aren't sure they picked the right one. The scroll is not the strategy. Finding the right posts to comment on is a research problem, and it has a repeatable solution.

B2B founders find the right LinkedIn posts to comment on by first identifying which influencers their ideal buyers actually follow, then monitoring those accounts for posts early in the engagement window (under 2 hours old), and filtering for posts that invite a substantive response rather than just validation. The goal is to place a sharp comment where your buyer is already paying attention, not where the post happens to be trending.

Why the feed is the wrong starting point

LinkedIn's feed is optimized for your engagement history, not for your pipeline goals. If you've been liking SaaS growth threads, you'll see more SaaS growth threads. That's fine for discovery, but it creates a structural problem: the posts surfacing in your feed are posts the algorithm thinks you'll engage with, which is not the same thing as posts your ideal buyers are reading.

The founders who build consistent inbound through commenting start from the buyer, not the feed. The question is not "what showed up in my feed today?" but "which people do my buyers follow, and what are those people posting right now?"

We've watched this pattern across dozens of account audits. Founders who comment on algorithmically surfaced posts build a vague sense of LinkedIn presence. Founders who comment on posts from influencers their buyers follow start getting DMs that open with "I keep seeing your name come up."

Step 1: build your target influencer list

Before you can find the right posts, you need a list of the right accounts to monitor. This is a one-time research investment that pays compound returns.

Start with 10-15 accounts in your buyers' orbit. These are not necessarily the biggest names in your category. They are the people your specific buyers follow and read. A few ways to find them:

Ask your customers directly. "Who do you follow on LinkedIn for [topic]?" gets you a list in one conversation. Most customers will name 3-5 people without hesitation. This is the most reliable signal you can get.

Look at who your best customers engage with. Go to a recently closed customer's LinkedIn profile. Click their activity. Look at what posts they've commented on or liked in the past 30 days. The authors of those posts are your targets.

Work the "also follows" signal. When you visit an influencer's profile, LinkedIn shows followers who overlap with people in your network. If multiple people you know follow the same account, that account is in your buyers' ecosystem.

Check who appears in your buyers' comment sections. When a customer posts something, look at who else is commenting. The senior people commenting on that post probably have overlapping audiences.

The list you're building here is not "big LinkedIn accounts in B2B SaaS." It's "the 12 people my CTO buyers in the 100-500 person company range actually read." That specificity is the entire point.

Step 2: filter for posts worth commenting on

Not every post from your target list is worth your time. Three criteria narrow it down:

Post age. LinkedIn's algorithm distributes new posts in waves. The first wave goes to your closest connections; if they engage, it goes wider. A sharp comment placed in the first 90 minutes gets seen by the early engagers, which are usually the most engaged people in that audience. A comment placed 18 hours later gets buried under 200 others and seen by almost no one.

The post invites a specific response. Posts that share a data point, a contrarian opinion, or a specific decision the author made are posts you can add something real to. Posts that are purely inspirational leave no surface for a substantive response. You want a post where you can say something a CTO or VP of Sales would find credible and specific, not a post where the only available response is agreement.

The post's audience matches your buyer. An influencer posting for developer audiences is different from the same influencer posting for GTM leaders, even on the same day. Look at the content of the post, not just the author. Comments on a post about sales compensation reach a different slice of an influencer's audience than comments on a post about engineering culture, even if it's the same influencer.

Step 3: write a comment that earns profile visits

This part is covered more thoroughly in our breakdown of how top LinkedIn creators drive engagement, but the short version: a comment that makes your buyer curious enough to click your name is worth twenty times a comment that gets a thumbs-up from the author.

What generates that curiosity: a specific data point from your own experience that adds to or complicates what the author said. A direct question that implies you've thought about this more than most people. A concise counterexample with enough specificity that it reads as something you've actually seen, not something you've read.

What doesn't: "Great post, fully agree." "This is so important." "We ran into this too, feel free to DM me." The algorithm treats these as low-quality engagement, and experienced readers skip them.

The comment is not a sales pitch. It's a demonstration of the thing you know. The profile visit comes from curiosity, and the curiosity comes from a comment that doesn't fully resolve the question it raises.

Step 4: track what's working

After 3-4 weeks of commenting on targeted posts, you should be able to answer three questions: Which influencers' audiences engage with my comments most? Which types of posts generate profile visits? Which comments led to connection requests or DMs?

If you can't answer these, you're optimizing blind. The founders who build real inbound from LinkedIn engagement treat it as a channel with measurable inputs and outputs, not a vibe-based activity.

The metrics worth tracking are not likes on your comments. They are: profile views in the 24 hours after a comment, connection requests from people you didn't already know, and (the lagging indicator) how many late-stage sales conversations include some version of "I've been seeing your name for a while." The earlier indicators -- profile views and connection requests -- tell you whether you're on track before deal signal arrives.

The influencer-selection mistake most founders make

When founders do start building a target list, they almost always over-index on follower count. The logic is straightforward: bigger audience means more exposure. In practice, the relationship between follower count and pipeline-relevant visibility is weaker than it looks.

A 400,000-follower account that posts for a general professional audience might have 200 CTO readers in your target company-size range. A 15,000-follower account focused on B2B SaaS operations might have 2,000 of them. A comment on the second account reaches a denser concentration of your buyers per impression.

We've looked at this pattern in account audits of founders who've built real inbound through LinkedIn. The accounts that generated the most inbound-attributable DMs were almost never the largest influencers they were commenting on. They were the mid-size accounts with tighter audience alignment. Our analysis of how Justin Welsh structures his LinkedIn strategy shows a version of this: his engagement rate is high precisely because his audience is self-selected around a specific topic rather than a generic professional interest.

Follower count is a proxy for reach. What you actually want is reach within your specific buyer cohort. Those are different things, and conflating them is how you end up spending 6 months building visibility with people who will never buy from you.

Making this executable in 10 minutes a day

The research above is a one-time investment. The daily habit is short: open your target influencer list, sort by recency, check for posts under 2 hours old that meet the criteria, write one or two sharp comments, and close the tab.

The whole execution is 10 minutes if the target list is built and the filter criteria are internalized. The part that takes 30 minutes is the scroll, which you've now replaced with a system. For a more detailed look at how to turn that presence into consistent inbound, our guide to LinkedIn inbound signals walks through the leading indicators worth watching and what to do when they spike.

<h2> <a name="faq" href="#faq" class="anchor"> </a> FAQ </h2> How many LinkedIn posts should a B2B founder comment on per day? Two to three well-placed comments on targeted posts outperform ten generic comments scattered across the feed. Quality and targeting matter far more than volume. Starting with two comments per day on posts from your monitored influencer list gives you enough data to see what's working within 3-4 weeks without burning time you don't have. How do I find which LinkedIn influencers my B2B buyers actually follow? The most reliable method is asking customers directly: 'Who do you follow on LinkedIn for [topic]?' takes one conversation and returns 3-5 names. The second method is checking the activity feed of recently closed customers to see whose posts they've been engaging with. Both approaches give you buyer-validated signal rather than guesses based on follower counts. How early do I need to comment on a LinkedIn post for it to get visibility? The first 60-90 minutes after a post goes live is when early engagers are most active and the algorithm is deciding how widely to distribute. Comments placed in this window get seen by the most engaged readers and appear near the top of the comment section before it gets crowded. After about 4-6 hours, you're largely invisible unless the post goes genuinely viral. What kind of LinkedIn comment actually drives inbound for B2B founders? Comments that add a specific data point, a direct counterexample, or a question that implies deeper familiarity with the topic generate more profile visits than comments that agree or validate. The comment should raise the reader's curiosity about who wrote it, not resolve everything on the spot. A comment that makes a senior buyer think 'I want to know more about this person' is doing its job. <hr> This article was first published on the <a href="https://www.getchime.co/blog/b2b-founders-linkedin-posts-comment?utm_source=devto&utm_medium=republish">Chime blog</a>. For more on LinkedIn pipeline tactics, <a href="https://www.getchime.co/blog?utm_source=devto&utm_medium=republish">visit the Chime blog</a>.

Using AI to support and defend your brand

Chime — Sun, 14 Jun 2026 03:04:13 +0000

AI systems are writing your brand's first impression. They are doing it right now, across ChatGPT, Perplexity, and every major AI search interface, and most B2B founders and senior leaders have no coherent plan to influence what gets said. That gap is worth closing before a competitor does.

AI tools pull from your entire content footprint -- forums, review sites, outdated press, social posts -- not just your owned properties. The most repeated claim surfaces, not the most accurate one. B2B brands that want to influence AI-generated summaries need to publish consistently, govern their messaging across every channel, and monitor what AI platforms are saying about them on a regular cadence. LinkedIn is one of the highest-signal surfaces for this work because it is both a publication channel and a citation source.

The first impression has been compressed

The old model gave brands time. A prospective customer would piece together an impression across multiple touchpoints: a press mention, a website visit, a product review, a conversation at a conference. Perception accumulated slowly, and there were many chances to correct a bad signal.

That model is largely gone for founders and companies with any public footprint. A buyer asks ChatGPT or Perplexity about your company, gets a two-paragraph summary, and walks away with a complete picture -- accurate or not -- before ever touching anything you control. The summary is not flagged as provisional. It does not come with a timestamp. It reads like the conclusion of research the buyer did not actually do.

What makes this genuinely hard is that AI does not prioritize your owned content. It pulls from whatever it can find: your website, third-party coverage, review platforms, LinkedIn activity, forum discussions, complaint threads. Volume often beats accuracy. A sustained stream of low-quality negative content can outweigh a smaller body of accurate positive content because AI systems weigh repetition heavily. Old positioning that you quietly moved away from three years ago can sit in the output alongside your current messaging, with nothing to distinguish them.

Your AI reputation is your entire content footprint, not just the parts you have invested in carefully.

Partial truths are harder to fix than false ones

Most B2B brands are not facing outright fabrication from AI systems. The more common problem is partial truths: accurate statements pulled out of context, outdated positions that were once correct, nuanced stances flattened into something that no longer reflects where you stand.

Partial truths are harder to dispute because there is something accurate in them. A competitor or a buyer who encounters a reductive summary of your positioning cannot be told it is simply wrong. And once an AI system has assembled a narrative from the sources it found, that narrative compounds. It gets reinforced every time someone asks a related question. It becomes what people know about you, and pushing it out requires more than publishing accurate content once. It requires replacing the sources the AI is drawing from -- consistently, over time, across multiple surfaces.

There is also a distribution effect worth taking seriously. AI-generated summaries get screenshotted and shared. Those shares become new inputs that reinforce the same narrative in future AI outputs. A misleading summary does not stay contained to the person who first encountered it.

What you can actually do about it

The practical playbook has three parts, and none of them are exotic.

Publish with enough frequency and specificity that the accurate version of your positioning is the most repeated version. AI systems surface what they see most. If your owned content is thin, outdated, or generic, the gap gets filled by whatever else is out there. The goal is not to produce a lot of content. It is to produce enough specific, on-point content that your actual positioning has more surface area than the distorted version.

LinkedIn is worth calling out specifically here. It is indexed by AI systems, it surfaces in citation pools, and posts stay in circulation long enough to compound. We have written about how LinkedIn content functions as a signal for AI search tools -- the short version is that consistent, specific activity on LinkedIn does more for your AI reputation than many founders expect. A post that clearly articulates your differentiated position is a citation waiting to happen.

Govern your messaging across every channel your brand touches. Inconsistent messaging does not get smoothed over by AI -- it gets amplified. If your website says one thing, your LinkedIn says another, and a two-year-old press release says a third, those three signals compete. AI systems do not resolve the contradiction in your favor. They surface whichever version appears most often or comes from a source they weight more heavily.

This is not about brand police work or enforcing tone guidelines. It is about making sure that the core claims you want associated with your brand -- what you do, who you do it for, what makes you different -- are stated consistently and often enough to dominate your content footprint.

Monitor what AI systems are actually saying about you. Most B2B founders have not done this. The exercise is simple: ask ChatGPT, Perplexity, and whatever AI search tool your buyers are likely to use what they know about your company. Ask specifically about your positioning, your competitors, your category. Read the outputs as if you are a buyer who knows nothing about you. Then ask whether the summary you just read would make you want to continue the conversation.

If the answer is no, or if the summary contains outdated framing you no longer stand behind, that is the signal. The response is not to file a correction request with the AI provider. It is to publish more and more specifically, so that the sources available to the AI system tell a better story.

Where LinkedIn fits in the brand defense stack

We work with B2B founders and senior leaders who are building inbound through their expertise on LinkedIn, and what we have noticed is that the AI reputation problem and the LinkedIn engagement problem are the same problem. Both require consistent, specific, public statements of your actual position. Both reward founders who show up with a point of view rather than distributing noise.

The founders in our network who are hardest to misrepresent by AI systems are the ones who have built a clear, specific, frequently repeated body of public content. They have said the same true things many times in different ways. When an AI system looks for signals about who they are and what they stand for, it finds a coherent answer because they have given it one.

The founders who are most vulnerable are the ones who post inconsistently, whose LinkedIn presence does not reflect their actual positioning, and who have let third-party content -- reviews, mentions, forum threads -- fill the space their own publishing left empty.

Founder-led LinkedIn activity that builds inbound pipeline is not separate from brand defense. It is brand defense. Every specific, well-articulated post is a citation in the pool AI systems draw from. Every time you clearly state your differentiated position in public, you are reducing the surface area for misrepresentation.

Every specific, well-articulated post is a citation in the pool AI systems draw from.

The governance question

There is a practical tension worth naming. Most B2B founders do not have a content governance process. They publish when they have something to say, respond to requests from PR or marketing teams, and let third-party content exist without auditing it. That approach worked reasonably well in a world where buyers formed impressions gradually. It works poorly in a world where a single AI-generated paragraph can substitute for a buyer's entire research process.

Governance does not have to be complex. The floor is: know what claims you want associated with your brand, publish those claims specifically and often enough to dominate your own content footprint, and check periodically what AI systems are saying about you so you know whether you are succeeding.

The question of what your brand content decision should actually be is worth thinking through carefully. Governance starts there, with clarity about what you actually stand for, not with a style guide or an approval workflow.

The time horizon is short

AI-generated first impressions are not a future problem. They are a current one. The brands that figure out how to influence what AI systems say about them in the next 12 months will have a structural advantage that is hard to replicate. The brands that wait are not pausing -- they are letting the current narrative calcify.

The entry point is not complicated: run the audit, find the gap between what AI systems say about you and what you actually want buyers to know, and close that gap by publishing the accurate version with enough frequency that it dominates. LinkedIn is the most accessible surface for that work. It is indexed, it compounds, and it does not require a content team or a PR agency to get started.

<h2> <a name="faq" href="#faq" class="anchor"> </a> FAQ </h2> How do AI systems decide what to say about my brand? AI systems pull from your entire public content footprint: your website, press coverage, LinkedIn activity, review platforms, forum discussions, and third-party mentions. They weight by repetition and source authority, not accuracy. The most frequently repeated claim tends to surface, which means thin or inconsistent owned content leaves room for third-party narratives to dominate the output. Can I correct what AI tools like ChatGPT or Perplexity say about my company? Not directly. There is no formal correction process that updates an AI system's outputs in real time. The practical approach is to publish accurate, specific content consistently enough that the sources available to the AI tell a better story. LinkedIn posts, updated website copy, and authoritative third-party coverage all feed into the citation pool these systems draw from. How often should B2B founders audit what AI systems say about their brand? A quarterly check is a reasonable floor. Ask ChatGPT, Perplexity, and any AI tools your buyers are likely to use what they know about your company, your positioning, and your category. Read the output as a buyer who knows nothing about you. If it contains outdated framing or misses your actual differentiation, that is the signal to publish more specifically and more often. Why does LinkedIn specifically matter for AI brand defense? LinkedIn content is indexed by AI search systems and appears in the citation pools these tools draw from when generating summaries. Consistent, specific activity on LinkedIn -- posts that clearly articulate your positioning and differentiation -- gives AI systems accurate source material to work with. It also compounds over time, meaning the more you publish, the harder it becomes for inaccurate third-party content to dominate your AI-generated reputation. <hr> This article was first published on the <a href="https://www.getchime.co/blog/using-ai-support-defend-brand?utm_source=devto&utm_medium=republish">Chime blog</a>. For more on LinkedIn pipeline tactics, <a href="https://www.getchime.co/blog?utm_source=devto&utm_medium=republish">visit the Chime blog</a>.

MIT's AI risk study: the public is exposed

Chime — Sat, 13 Jun 2026 03:03:48 +0000

MIT FutureTech just published the largest expert survey on AI risk we have seen: 272 respondents, 24 distinct risks, weighted by domain expertise. The finding that stood out to us is not about rogue models or AGI. It is about accountability: the people most exposed to AI downsides have the fewest tools to do anything about it.

The MIT FutureTech study found that ordinary users face the highest AI risk concentration while having almost no power to audit, challenge, or opt out of the systems shaping their outcomes. Experts ranked misuse by developers and unaccountable deployment far above speculative long-horizon risks like AGI. For B2B operators, this is a trust problem that lives inside the customer relationship, not inside the model.

What the study actually measured

The survey did not ask experts to speculate about 2040. It asked them to score 24 near-term and medium-term risks across two dimensions: likelihood and severity. Responses were weighted by each respondent's stated domain expertise.

The risks that ranked highest were not the ones that get mainstream coverage. Misinformation amplification, biased decision systems in hiring and credit, and opaque data practices all outscored the model-capability fears that tend to dominate tech press cycles. The through-line connecting the top-ranked risks is a structural asymmetry: developers hold information about how systems work; users don't.

Why this lands differently for B2B founders

When you sell a hiring tool, a credit underwriting model, a lead-scoring system, or an AI-assisted advisory service to another business, your buyer's customers sit at the bottom of the accountability stack. They interact with outputs from a system they cannot inspect, built by a team they have never spoken to, sold through a vendor relationship they are not party to. If the model produces a biased recommendation, a factually wrong answer, or a quietly degraded output, the person who bears that cost is not your enterprise buyer. It is the end user your buyer serves.

The experts MIT surveyed were explicit about where responsibility sits in that chain: with the developer, not the deployer, and certainly not the user. That is a meaningful gap from where legal liability currently lands.

The trust problem is already showing up in pipeline

We work with a range of B2B founders and senior operators at services and SaaS companies. The MIT framing matches something we have been watching in sales cycles over the past 12 months: buyers are asking harder questions about AI components than they were in 2023. Not "does this use AI?" but "how does it make decisions?" and "what happens when it's wrong?"

The operators who have good answers to those questions are closing faster. The ones who treat the AI stack as a differentiator to showcase rather than a system to be accountable for are getting more objections, longer security reviews, and more ghosting at the evaluation stage.

Buyers are more skeptical about AI claims than they were 18 months ago, and that skepticism is visible in how they engage with content and how they structure vendor evaluations. They are internalizing the MIT study's conclusion before the study exists for them, because they have been burned or they know someone who has.

Four things the study implies for how you operate

Disclose more than you are required to. The study found that opaque data practices and unexplained automated decisions were among the highest-ranked risks. Voluntary transparency, spelled out in plain language in your product documentation and sales process, is the cheapest form of risk mitigation available.

Build an audit trail users can actually read. Not a log file. A readable account of what the system did, why, and what the confidence level was. Most AI products do not have this. The ones that do are already using it as a sales differentiator.

Separate the AI component from the outcome. When something goes wrong, users need a way to distinguish between "the AI recommended X" and "the product decided X." That distinction matters for trust, for correction, and for any future regulatory environment that asks who made the call.

Treat your end user's exposure as your problem, not your buyer's. Your enterprise buyer will sign the contract. The user at the bottom of the stack has no contract with you and no recourse through normal channels. The MIT experts are saying that moral responsibility for that user's exposure sits with the developer. Building as if that is true, before regulation requires it, is what earns durable buyer trust. In the audits we run, it is already separating operators who close from operators who stall.

What the study does not say

This is an expert survey, not a controlled study, and the respondents skew toward safety and policy circles, worth noting when calibrating. What it does establish cleanly: the experts closest to AI development do not think the biggest risks are science fiction scenarios. They think the biggest risks are accountability gaps that exist right now, in production, in products that are already shipped. That is a more useful finding than most of what gets published on AI risk.

The LinkedIn angle

The operators building the most credible presence right now are the ones willing to say what their AI stack can and cannot do. Posts that engage honestly with AI limitations generate more substantive conversation than posts that announce AI as a feature. The comment sections on those posts look different: more questions, more replies, more people tagging colleagues.

Buyers are more skeptical about AI claims than they were 18 months ago, and that skepticism is visible in how they engage with content. The founders who acknowledge that skepticism directly, and back it up with specifics, are getting engagement from the exact audience they want to reach.

For a deeper look at how AI is changing the distribution math for B2B content, our piece on LinkedIn as a signal source for AI search covers the compounding effects that expert-credibility posts generate over time. And if the trust problem in AI sales cycles connects to what you are seeing in your own GTM motion, the AI strategy trust problem piece gets into the mechanics of why the credibility gap is widening.

The study tells you where expert consensus says the accountability gap sits right now, in production, in shipped products. Acting on that before it becomes a compliance requirement is the cleaner path.

<h2> <a name="faq" href="#faq" class="anchor"> </a> FAQ </h2> What were the top-ranked risks in the MIT FutureTech AI risk study? The MIT FutureTech survey of 272 AI experts ranked misinformation amplification, biased automated decision-making in hiring and credit, and opaque data practices among the highest risks. Speculative long-horizon risks like AGI scored lower. The common thread across the top-ranked risks was accountability asymmetry: developers hold information that affected users cannot access or challenge. Why does the MIT AI risk study matter for B2B SaaS founders specifically? B2B SaaS founders often sit in the middle of an accountability chain: they build the AI system, sell it to an enterprise buyer, and the enterprise buyer's customers bear the downstream risk. The MIT experts placed moral responsibility for that exposure with the developer, not the deployer or the end user. That creates a gap between current legal liability frameworks and where expert consensus says accountability belongs — a gap that is already showing up in enterprise sales cycles as harder buyer questions about AI decision-making. How should a B2B founder respond to growing AI risk scrutiny from buyers? The most effective response we have seen in practice is voluntary transparency: plain-language documentation of how AI components make decisions, what confidence levels look like, and what happens when the system is wrong. Founders who can answer 'how does it make decisions?' and 'what is the correction process?' in a sales conversation are closing faster than those who treat AI as a marketing differentiator and stop there. What does the MIT AI risk study say about who is responsible when AI causes harm? The study's expert consensus placed primary responsibility with developers, not with deployers or users. The reasoning is structural: developers have the most information about how a system works and the most power to design safeguards, while the people most exposed to harm typically have neither the technical access nor the contractual standing to push back. The study does not make legal claims, but it reflects where expert moral accountability thinking is concentrated. <hr> This article was first published on the <a href="https://www.getchime.co/blog/mit-ai-risk-study-public-exposed?utm_source=devto&utm_medium=republish">Chime blog</a>. For more on LinkedIn pipeline tactics, <a href="https://www.getchime.co/blog?utm_source=devto&utm_medium=republish">visit the Chime blog</a>.

LLMs are picking winners in your niche

Chime — Sat, 13 Jun 2026 03:03:47 +0000

PostHog published something worth reading last month. Their traffic from LLM referrals grew 41x in two years, and it converts better than almost any other source they have. The uncomfortable corollary: not all of their products get recommended at the same rate. Some are well-known to the models; others are effectively invisible. We think the same asymmetry is playing out right now across B2B founders and operators, and the gap is widening.

LLMs recommend people and products based on the citation inventory they were trained on and the sources they can retrieve in real time. For B2B founders and operators, LinkedIn is one of the highest-signal surfaces feeding that inventory. Operators with consistent, specific, citable content on LinkedIn are far more likely to show up when a buyer asks an AI tool "who should I talk to about X." Those without it are not. The window to build that presence before recommendations calcify around a short list of recognizable names is shorter than most people think.

How LLMs decide who to recommend

When someone asks ChatGPT, Claude, or Perplexity "who is the best fractional CMO for B2B SaaS" or "which consultant should I hire for enterprise pricing strategy," the model is not running a Google search and surfacing whoever has the most backlinks. It is drawing on two things: what was in its training data, and what it can retrieve right now from high-authority sources.

Training data is the longer game. It reflects what has been written about you, what you have written, and where that writing lives. Retrieval is the shorter game. Perplexity and ChatGPT's browsing mode pull live content from sources they trust. LinkedIn is one of them. So is your company blog. So are publications that quote you.

The PostHog team found that their best-known products get recommended consistently, while newer or less-documented products get skipped entirely. The models know what they know. If you have not created a clear, consistent, retrievable body of work that answers the questions your buyers are asking, the model will recommend someone who has.

This is not a future problem. Across the founders and operators we work with, we are already seeing inbound come in with the note "an AI tool suggested I reach out to you." That was rare 18 months ago. It is no longer rare.

Why LinkedIn specifically

There are three reasons LinkedIn matters more than most operators expect for LLM visibility.

First, LinkedIn posts are indexed and retrievable. Perplexity pulls them. ChatGPT's browsing mode pulls them. When a model is trying to answer "who are the credible voices on go-to-market for vertical SaaS," it is pulling from content that exists in retrievable form, and LinkedIn posts meet that bar in a way that a Slack message or a podcast appearance does not.

Second, LinkedIn has a trust signal that general web content lacks. The platform attaches professional identity to content. A post about pricing strategy written by someone whose profile says "VP of Product at three Series B companies" carries a different signal than an anonymous blog post. Models that care about source authority register that.

Third, volume creates citation density. One good post is not enough. The operators who show up in AI recommendations tend to have published consistently on a narrow set of topics over a meaningful period of time. The models see a coherent body of work and treat it as authoritative. One-off posts, even excellent ones, do not build that.

We wrote about how LinkedIn has become a source signal for AI search tools in more depth here, and the same dynamics we described then are accelerating now.

The citation inventory problem

Here is the practical issue: most B2B founders have not built a citation inventory on LinkedIn. They have posted occasionally, usually about company news or surface-level commentary on trends. When we run content audits, we consistently find that founders have expertise that is nowhere in their public record.

A founder who has spent five years building and selling to mid-market professional services firms knows specific things: how those buyers evaluate, what objections they raise at what stage, which integrations matter and which do not. That knowledge is an asset. If it only lives in sales calls and private Slack channels, it does not exist for the purposes of LLM recommendations.

The fix is not complicated, but it is consistent. You need a body of public work that answers the questions your buyers are already asking AI tools. That means publishing on LinkedIn with enough specificity that a model ingesting your content can accurately describe what you do and who you are for.

Vague is the enemy here. "I help founders scale" is not citable. "In eight engagements with Series A SaaS companies, the CAC problem turned out to be a segmentation problem four times" is citable. The model can do something with the second sentence. It cannot do anything with the first.

What the models are actually trained to look for

AEO (answer engine optimization) is the emerging field PostHog references. We do not love the term because it has already attracted the same keyword-stuffing instinct that ruined early SEO. But the underlying principle is sound: if you want to be recommended by AI tools, you need to write content that answers specific questions accurately, in depth, in a format that is easy to cite.

That translates to a few practical things on LinkedIn.

Specificity beats breadth. A founder who has 40 posts about pricing strategy for B2B SaaS will be recognized by models as an expert on that topic. A founder who has 40 posts across 15 topics will be recognized as someone who posts a lot. Own a lane.

First-person data beats general commentary. "According to Gartner..." is a worse citation anchor than "across the 12 companies I've worked with in this segment." Models are trained on a lot of Gartner paraphrasing. They are trained on less specific practitioner data. First-person specificity stands out.

Frequency creates the training signal. PostHog's 41x growth in LLM traffic did not come from one good piece of content. It came from a consistent volume of useful, citable writing over two years. The same principle applies to individuals. A single excellent post in March will not make you a recommended expert in November. Twelve good posts per month for twelve months might.

We covered the compounding logic of consistent LinkedIn presence in detail in our piece on how Google's no-click data changes the distribution math. The argument there applies with even more force to AI-driven recommendations, where there is no second page of results.

The short list problem

Here is what makes this urgent. LLMs do not return a long tail of results the way Google does. When a buyer asks an AI tool for a recommendation, they get three to five names, occasionally fewer. The model is not going to say "here are 47 options, sorted by relevance." It is going to say "based on what I know, these are the people worth talking to."

That is a radically different competitive dynamic than traditional search. In traditional search, being on page two is mediocre but not catastrophic. In AI search, not being in the top three to five is functionally invisible. The buyer closes the chat, contacts whoever was recommended, and never thinks about anyone else.

The short list is already forming in most niches. Not because one player has locked it up, but because a small number of operators have built enough of a citation inventory that the models default to them. The operators who are not on that list yet have a narrowing window to build their way onto it.

This is not hypothetical. Run the prompt yourself. Open ChatGPT or Claude, turn off memory, and ask: "who are the best consultants for [your specific niche]?" See who shows up. If you are not in the results, ask a follow-up: "what would I search to find someone who does what [your name] does?" If the model does not know your name, that is your baseline.

What to do this week

The practical starting point is the same one PostHog recommends for products, applied to people.

Run three to five prompts that your buyers would realistically use. "Best [your role] for [your niche]." "Who should I hire if I need help with [your core problem]." "Recommended [your category] consultant." Do this across ChatGPT, Claude, and Gemini. Note where you appear, where you are missing, and where you are misrepresented.

Then look at your LinkedIn content from the last six months and ask honestly: does this add up to a citable body of work on a specific topic? If the answer is no, you know what to build.

The good news is that the bar for individual operators is lower than it is for products. PostHog is competing against dozens of well-funded, well-documented analytics platforms. You are probably competing against a handful of operators who have not yet realized this game is being played. Most of them are posting sporadically about general trends. You do not need to outpublish them. You need to out-specify them.

Forty posts over six months, all on the same specific problem you solve, written with enough first-person data and practitioner detail that a model can accurately describe what you do. That is the investment. The return is showing up on a short list that your buyers consult before they ever open a search engine.

The models are not going to return a long tail of results. They are going to return three to five names. That short list is forming now.

The founders we see winning on LinkedIn right now are not the ones with the most followers. They are the ones building the kind of public record that makes a model confident enough to recommend them by name. That is a different game than follower growth, and it is the one worth playing.

<h2> <a name="faq" href="#faq" class="anchor"> </a> FAQ </h2> How do LLMs decide which experts to recommend when someone asks for a consultant or specialist? LLMs draw on two things: content from their training data and real-time retrieval from trusted sources. For individual experts, that means the models look at what you have written publicly, where it is hosted, and how consistently you have addressed a specific topic. LinkedIn posts, blog articles, and quotes in publications all feed that inventory. Operators with a consistent body of work on a narrow topic are more likely to be recommended than those who post occasionally across many topics. Does LinkedIn content actually show up in AI search tools like Perplexity or ChatGPT? Yes. Perplexity indexes LinkedIn posts directly, and ChatGPT's browsing mode can retrieve public LinkedIn content. The platform's combination of professional identity signals and indexed content makes it a higher-trust source for AI retrieval than most general web content. A post about enterprise pricing strategy attributed to a credible professional profile carries more citation weight than an anonymous blog post on the same topic. What is answer engine optimization (AEO) and how is it different from SEO? AEO is the practice of optimizing content to appear in AI-generated answers rather than traditional search results pages. Where SEO focuses on ranking factors like backlinks and keyword density, AEO prioritizes creating specific, accurate, citable content that models can confidently use when answering a user's question. For B2B operators, this means writing with enough first-person specificity and topical depth that a model can accurately describe your expertise and recommend you to a relevant buyer. How many LinkedIn posts do I need before LLMs start recommending me? There is no precise threshold, but the pattern we observe is that consistency over time matters more than any single post. Operators who publish 8 to 12 times per month on a specific topic for six months or more tend to build enough citation density for models to recognize them as an authority in that area. One excellent post does not accomplish this. A coherent body of work on a narrow problem does. <hr> This article was first published on the <a href="https://www.getchime.co/blog/llms-picking-winners-linkedin?utm_source=devto&utm_medium=republish">Chime blog</a>. For more on LinkedIn pipeline tactics, <a href="https://www.getchime.co/blog?utm_source=devto&utm_medium=republish">visit the Chime blog</a>.

Kill the GTM stack: Aurasell CEO's case

Chime — Fri, 12 Jun 2026 03:03:51 +0000

At SaaStr AI 2026, Aurasell co-founder and CEO Jason Eubanks skipped the AI futurism. He put the exact go-to-market stack he ran at his last company on screen, told the room what it cost, and made a specific argument for why the whole model is breaking. We pulled the key data points from his presentation because they're the kind of numbers that make the abstract real.

Jason Eubanks' GTM stack at Harness ran 22 products, cost over $3M per year in software fees, and required 11 ops people just to maintain it. His argument: adding AI agents to that fragmented stack makes the problem worse, not better. The fix is a unified data layer first, then automation built on top of it.

The 24-30% problem nobody measures

Eubanks opened with a number most revenue leaders don't track: B2B sellers spend 24-30% of their time in front of prospects and customers. That's it. The other 70-plus percent goes to context switching, manual account research, prep, follow-up, and internal overhead like deal reviews and QBR prep.

None of that is selling. All of it is work you're paying full quota-carrying salaries to perform.

His challenge to the room was direct: if you can't say what that percentage is for each of your reps today, go measure it. He treats selling time as a top-line KPI, not an HR metric. That framing matters. When you treat it as a productivity metric rather than a morale metric, you stop managing it with motivation and start managing it with structure.

What the legacy stack actually costs

Eubanks didn't theorize. He put his old numbers on the screen from his time at Harness.

The stack ran 22 products. It cost $3M or more per year in software fees. It required 11 ops team members just to keep it standing. Those 11 people weren't driving revenue. They were stitching together integrations, patching workflow layers that conflicted with each other, and reconciling data across multiple databases. The one output they actually wanted, a single view of the customer journey, stayed permanently out of reach.

The deeper audit came through an exercise he called Project X-Ray. Mid-COVID, his board asked him to cut burn and accept slower growth. So he logged every activity, every tool, and every overlap across the org. The finding that stuck: reps were working inside 10 to 12 products a day to do their jobs, bleeding time to context switching the entire way.

That's the math of tool sprawl when it's concrete. It's not an abstract efficiency concern. It's 11 salaries, $3M in fees, and reps spending three-quarters of their day not selling.

Why adding agents to this makes it worse

Most legacy vendors are responding to AI pressure by bolting agents onto their existing product. Eubanks named why that backfires.

Every niche tool in a typical GTM stack carries its own siloed database. That silo might sync with your CRM at the field level, which looks clean on a data map. But the context -- the actual conversations, activities, and signals -- stays trapped inside each silo. That context is what an agent needs to act intelligently. Without it, the agent is guessing.

When legacy vendors layer agents onto fragmented data, you get agents running on a fraction of the relevant metadata, blind to the full picture. Eubanks calls the result "agentic thrash": low-quality automation at best, and at worst agents that act autonomously and step over each other. Adding more agents to a fragmented stack doesn't fix sprawl. It compounds it, and drives costs up while doing so.

This is the part of the AI sales pitch most vendors are quietly avoiding. The capability of the agent is only as good as the data feeding it. If the data layer is a patchwork of 22 tools, the agents inherit all the gaps.

The architecture Aurasell is betting on

Aurasell's architecture starts from the data, not the agents. Three layers:

A unified data foundation. Structured and unstructured data in one place. The platform ships with 900M contacts and 85M accounts, auto-enriched, with room to extend through custom enrichment.

A conversational context layer. Every conversation, every channel, every signal, feeding one context graph instead of a dozen silos.

An automation layer on top. Some agents come prebuilt and run autonomously. Others you build in natural language. The coverage is described as contact to contract -- the full sales process for both your team and the buyer.

The deployment model is the part worth watching for adoption. You can run Aurasell as your AI-native CRM and migrate off your existing tools. Or you can run it on top of Salesforce or HubSpot as an intelligence layer. That second option lowers the switching cost for teams that can't rip out their CRM this quarter, which is most of them.

What this means for operators building now

The Eubanks presentation is useful even if you never buy Aurasell. The Project X-Ray exercise is reproducible. Log every tool your team touches in a week. Log the overlaps. Count the ops headcount required to maintain the integrations. Then ask whether the productivity gains from each tool justify what it actually costs when you include the maintenance burden and the context-switching tax on your reps.

Most operators who run this exercise find the answer is no for a meaningful portion of their stack. The 22-tool number at Harness isn't an outlier. We've seen similar patterns across founders we work with who are building lean GTM motions and realize their tooling overhead is quietly eating the margin they thought they had.

The AI-native CRM argument Eubanks is making is a bet that the platform layer is about to consolidate the way the productivity suite did. Whether Aurasell wins that race or someone else does, the underlying logic holds: agents built on fragmented data don't solve the problem. They accelerate it.

For operators building inbound through LinkedIn and content, the parallel is worth sitting with. If your distribution stack has the same fragmentation problem -- different tools for scheduling, analytics, engagement tracking, and outreach that don't share context -- you're running a version of the same tax Eubanks described. The 24-30% selling-time number applies to content operators too. Most of the time goes to finding the right posts, writing the right comments, and figuring out what's working. The part that's actually generating pipeline is smaller than it looks.

That's the operational problem Chime is designed around. Not posting more. Not more tools. A tighter loop between the right engagement and the right context, so the time you do spend on LinkedIn is the 24-30% that actually moves the needle, not the overhead that surrounds it.

The GTM stack audit Eubanks ran at Harness is the kind of exercise that gets uncomfortable fast. The patterns across LinkedIn's top creators show a similar consolidation happening on the content side: the operators generating real inbound aren't running more tools, they're running fewer with more intentional distribution.

FAQ

What is GTM stack consolidation and why does it matter for sales teams?

GTM stack consolidation means reducing the number of separate tools your revenue team uses to do their job, and replacing fragmented point solutions with a unified platform that shares a single data layer. It matters because tool sprawl forces reps to context-switch constantly, which reduces the time they spend actually selling. Aurasell CEO Jason Eubanks found that reps at Harness were working inside 10 to 12 products a day, and the overall stack required 11 ops people and $3M per year just to maintain. Consolidation reduces that overhead and gives AI agents the unified context they need to work effectively.

How much time do B2B sales reps actually spend selling?

According to data Eubanks presented at SaaStr AI 2026, the average B2B seller spends just 24 to 30 percent of their time in front of prospects and customers. The remaining 70-plus percent goes to manual research, prep, follow-up, deal reviews, and context switching between tools. Eubanks argues this number should be tracked as a top-line KPI rather than a soft productivity metric, because it directly limits the output ceiling of any GTM org regardless of headcount.

Why do AI agents fail when added to a fragmented GTM stack?

AI agents require unified context to act intelligently. In a fragmented stack, each tool carries its own siloed database. Even if those tools sync at the field level with a CRM, the conversational context -- the actual signals, activities, and interactions -- stays trapped inside each silo. Agents built on that fragmented data are working with incomplete information, which produces what Eubanks calls 'agentic thrash': low-quality automation where agents can act autonomously and conflict with each other. Adding more agents to a fragmented stack compounds the problem rather than solving it.

What is Aurasell's approach to replacing the legacy GTM stack?

Aurasell is built on three layers: a unified data foundation with 900M contacts and 85M accounts, a conversational context layer that pulls every signal from every channel into one graph, and an automation layer where agents run on top of that complete data set. The platform can replace existing CRM tools entirely or run on top of Salesforce and HubSpot as an intelligence layer, which lowers the switching cost for teams that can't migrate their CRM immediately. The core bet is that agents built on a unified data layer outperform agents bolted onto fragmented point solutions.

This article was first published on the Chime blog. For more on LinkedIn pipeline tactics, visit the Chime blog.

Figma's CMO on safe marketing being risky

Chime — Fri, 12 Jun 2026 03:03:50 +0000

First Round's Executive Function podcast doesn't always produce something worth flagging. This episode with Sheila Joglekar Vashee, Figma's CMO, does. The framework she uses to think about brand risk is directly transferable to operators building visibility at a fraction of Figma's scale.

Sheila Joglekar Vashee argues that playing it safe in marketing is itself a risk. The conservative move -- repeating what worked last quarter -- compounds into brand irrelevance faster than a single bad moonshot would. Her framework: run marketing as a portfolio, with maintenance bets keeping the business stable and creative bets taking the shots that produce step-change outcomes.

"Ubiquity is the opposite of cool"

Vashee picked up this line from Urban Outfitters' CEO during her retail years. Gap in the 1990s is her case study. Gap was everywhere, and then it wasn't interesting. The distribution won. The brand lost.

She points to Harley Davidson and Apple as the counterexamples -- companies that scaled to mass reach without becoming generic. What those brands held onto is harder to name than it is to recognize. She calls it "the challenger spirit." The sense that the brand still has something to prove, even when it clearly doesn't need to.

For Figma, the challenge is real. The product is widely used. It's in design teams at companies of every size. Widespread usage and beloved brand are not the same thing, and Vashee knows it.

The pressure this creates on marketing is specific: growth can't come from ubiquity plays. Vashee is explicitly wary of the kind of growth tactics that win impressions at the cost of perception.

The spammy TikTok ad problem

Her framing here is worth quoting directly: "Think about any spammy ad you've seen on TikTok. That doesn't improve brand perception, but it gets your attention in the moment and it might make you click. Over time, that's detrimental to a company's brand, even if it's effective as a local maximum for that channel."

Local maximum is the phrase to hold. A local maximum in paid social -- high CTR, decent CPL, good short-run numbers -- can coexist with a long-run brand penalty that doesn't show up in the dashboard for 18 months. By the time it shows, the brand association is set.

For operators building on LinkedIn, the equivalent pattern is the engagement-bait comment. It gets a reaction in the moment. It might pull a follow. Over time, if that's all you're doing, the signal you send to the people in those comment sections is that you're optimizing for attention, not contributing to the conversation. Those are different things, and audiences eventually tell them apart.

Marketing as a portfolio

The framework Vashee uses to hold this tension: treat marketing spend like an investment portfolio, not a single bet.

Maintenance bets are the predictable plays that keep revenue coming in and the funnel moving. They're not exciting. They're not supposed to be. They exist so the business doesn't miss its numbers while the creative bets are running.

Moonshots are the ideas that "always seem crazy when you look at them individually." The ones that, if you had to defend each one in isolation in a budget meeting, would probably get cut. In a portfolio, you can hold them. Because if one fails, the maintenance layer absorbs it.

This is a different mental model than most operators use. The default is to ask "will this work?" about each tactic in isolation. The portfolio model asks "what mix of bets, across what range of certainty, gives us the best chance of a step-change outcome without torching the business if the moonshots miss?"

The practical consequence: she runs campaigns she expects to fail. Not casually, and not as an excuse for lazy thinking, but as a deliberate structural choice. The creative breakthrough doesn't come from batting 1.000. It comes from taking enough swings that one of them connects in a way the cautious version never would.

Why AI needs optimistic stories

The third thread in the episode is less tactical but worth noting. Vashee thinks the AI moment is being underserved by pessimistic narratives.

Her argument: every major platform shift -- the internet, mobile, social -- created enormous opportunities. The coverage of those moments was not uniformly positive at the time, but the builders found the optimistic frame and ran with it. AI is getting more fear-of-displacement coverage than it is getting joy-of-building coverage.

She sees room for brands, including Figma, to recapture the energy of the early-builder moment. The joy of making things, which is pretty close to what Figma's product enables, is a more durable positioning than "don't get left behind."

For operators whose content strategy leans heavily on the urgency angle -- the "AI is coming for your job" type of hook -- this is worth sitting with. Urgency gets clicks. Optimism about what's buildable now is a harder frame to establish and a stickier one once it's there. The creators we track who build real inbound over time tend to sit in that second category. Not because they ignore risks, but because their audience shows up to learn what's possible, not to be warned about what isn't.

What this means for operators at a smaller scale

Vashee is running marketing for a company heading toward a public market. The absolute numbers don't translate. The logic does.

The portfolio model applies whether you have $50,000 in marketing budget or $50 million. Operators building visibility through LinkedIn engagement are, in most cases, running a maintenance-only strategy: showing up consistently, commenting on the same influencers, repeating the format that worked last month. That's fine as a maintenance layer. The operators who break through tend to have one or two creative bets running in parallel -- a format they haven't tried, an audience they haven't engaged with, a perspective they've been sitting on because it's too contrarian to feel safe.

The portfolio frame makes the creative bet feel less reckless. You're not gambling the business on one campaign. You're allocating a slice of capacity to it, running it alongside the reliable layer, and learning from whatever happens.

On the ubiquity point: the LinkedIn equivalent of "ubiquity kills cool" is the comment that sounds like everyone else's comment. If your comment on a popular post reads the same as the 15 comments above it, you've bought yourself a local maximum -- maybe a few profile views -- at the cost of actual signal. The operators who build real inbound from LinkedIn engagement are not the ones posting most frequently, and they're not the ones saying the safest thing. They're the ones with a point of view that takes a small risk.

That's Vashee's argument about marketing at Figma. It scales down farther than she probably intended.

You can listen to the full episode on YouTube, Apple Podcasts, or Spotify.

<h2> <a name="faq" href="#faq" class="anchor"> </a> FAQ </h2> What does Figma's CMO mean by running marketing as a portfolio? Sheila Joglekar Vashee divides marketing activity into maintenance bets (predictable plays that keep the business stable and the funnel moving) and moonshot bets (creative ideas that seem risky in isolation but are manageable within a broader portfolio). The point is that you need the maintenance layer to absorb the failures from the moonshots, and you need the moonshots to produce the step-change outcomes that maintenance alone never will. Why does Vashee say safe marketing is risky? Repeating the conservative play -- the tactic that worked last quarter, the format everyone else is using -- compounds into brand irrelevance over time. A single failed moonshot is recoverable. Years of playing it safe can erode the brand positioning that makes a company worth caring about in the first place. The risk isn't visible in any single quarter, which is exactly why it's dangerous. What is the 'local maximum' problem in marketing? A local maximum is a channel or tactic that performs well on short-run metrics -- click-through rate, cost per lead, impressions -- while quietly damaging the brand's longer-term perception. Vashee's example is spammy TikTok ads: effective at getting clicks, detrimental to brand trust over 12 to 18 months. The dashboard looks fine until it doesn't. How does the 'ubiquity is the opposite of cool' principle apply to LinkedIn content strategy? Vashee's point about Gap in the 1990s applies directly to comment strategy: when your comment reads the same as the 15 comments above it on a popular post, you've optimized for presence at the expense of signal. The operators who build real inbound from LinkedIn engagement tend to take small risks with their point of view rather than repeating the consensus. Distinctiveness is what earns the profile visit. <hr> This article was first published on the <a href="https://www.getchime.co/blog/figma-cmo-safe-marketing-risky?utm_source=devto&utm_medium=republish">Chime blog</a>. For more on LinkedIn pipeline tactics, <a href="https://www.getchime.co/blog?utm_source=devto&utm_medium=republish">visit the Chime blog</a>.

Life of leverage: what Eric Jorgenson gets right

Chime — Thu, 11 Jun 2026 03:04:25 +0000

We went through Eric Jorgenson's Life of Leverage course to pull out what actually applies to operators building inbound pipeline. The material extends Naval's four leverage types further than most summaries let on, and one framing in particular changes how you assess your own working hours.

Eric Jorgenson's Life of Leverage course extends Naval Ravikant's four leverage types (people, capital, code, and media) into a practical operating framework. The core argument: in the current environment, 8 hours of input can produce 800 hours of output if you choose your tools and systems correctly. The course teaches how to identify which activities have leverage potential and which are trapping you in linear time-for-money exchanges.

The chainsaw problem

If you have 8 hours and an axe, you produce 8 hours of wood-chopping output. Give that person a chainsaw, and you get 80 hours of equivalent output. A tractor pushes that to 800. Five employees with tractors gets you to 4,000 hours of output while you sit elsewhere.

Most operators default to the familiar tool without checking whether a higher-leverage option exists.

What makes this framework useful is how it reframes the question you ask at the start of each day. The question stops being "what will I work on today?" and starts being "what tool am I using to do this work, and does the tool match the stakes of the task?"

The four types of leverage, extended

Naval's four types are the starting point. Jorgenson's course layers on how they compound and, more usefully, how to tell whether you're actually using any of them.

People. Ray Dalio reportedly runs a 50:1 labor ratio: for every hour he spends in a room with someone, they spend 50 hours executing. The question for operators isn't "do I have a team?" but "does my involvement in each project produce multiples of itself in output from others?"

Capital. Capital is leverage most founders think about in terms of funding. Jorgenson pushes it into everyday operations: return on owned assets, cost structure, income per hour not worked. Those last two metrics rarely appear on dashboards, but they track something real.

Code. The one type with zero marginal cost to replicate. Once built, software runs for free at any scale. Operators who aren't building or using software assets are competing at a structural disadvantage.

Media. The example that makes this concrete: a single tweet written over an afternoon in 2020 reached over a million impressions. On that same day, a college football stadium held 60,000 people. Nearly 20 stadiums worth of people had seen one piece of content produced once.

For operators building LinkedIn presence, media is the most accessible type and the one most people underuse. A well-placed comment in a high-traffic thread reaches the commenter's network, the post author's audience, and the thread itself for as long as it stays active. That's a single input with a long, non-linear tail.

How to tell if you're actually using leverage

Jorgenson offers seven diagnostic questions. We're pulling them here because they function as a quick audit for any operator:

Can you quickly sort through options to find high-leverage solutions?
Can you solve problems in permanent or semi-permanent ways, rather than fixing the same issue repeatedly?
Can you instinctively ignore low-leverage opportunities without needing to reason through them?
Are you confident that today's moves are part of a longer-term plan?
Have you developed sensitivity to time-wasting, the kind that triggers a reflex rather than a judgment call?
Can you deploy resources confidently?
Have you redefined what success looks like for your specific situation?

These questions work because they locate leverage as a skill, not a circumstance. Most operators assume leverage is something you get when you're big enough to have a team or a large following. Jorgenson argues it's a capacity you build, and the questions tell you where your training gaps are.

The metrics that track leverage are similarly useful: percentage of your time spent on work only you can do, income per hour worked versus income per hour not worked, revenue per employee. If you have not checked those numbers recently, that's a signal in itself.

What this means for LinkedIn specifically

Most operators treat LinkedIn as linear: post, get views, repeat. The higher-leverage version is building a system where each input compounds: a comment that reaches the post author's 40,000 followers, not just your own 800 connections. A relationship built through consistent presence in the right threads produces referrals, inbound conversations, and introductions you didn't have to engineer separately.

The founders we work with who build inbound fastest are almost always the ones who have identified two or three high-traffic posts per week from the right creators and engage with genuine, specific responses. The engagement isn't the output; it's the tool. The output is visibility in front of the exact audience they're trying to reach.

For more on how this plays out in practice, the LinkedIn top creators patterns breakdown shows how sustained engagement compounds over a 13-week window, and the goal gradient hypothesis on LinkedIn piece explains why proximity to a visible outcome changes how audiences respond to you.

The metric most operators ignore

Jorgenson's metric list includes one that almost no operator tracks: income per hour not worked.

Income per hour not worked measures how much your systems are generating while you're doing something else. A course you built and sell asynchronously has a high value on this metric. A service business where you personally deliver every engagement has a value of zero.

For LinkedIn operators, this metric translates to a question about what your content and presence are doing while you sleep. If a post from three months ago is still pulling people into your profile, that's income-per-hour-not-worked in its LinkedIn form. If your only discovery mechanism is posting today and hoping today's algorithm serves it today, you're at zero.

The audits we run for founders regularly show that older posts, especially ones that sparked genuine comment threads, continue to generate profile visits long after the initial distribution window. Operators who know this build for it deliberately. Operators who don't are getting one-time returns on assets that could keep working.

The $300 question

A $300 input that restructures how you think about your working hours has an ROI that is genuinely hard to calculate.

<h2> <a name="faq" href="#faq" class="anchor"> </a> FAQ </h2> What is Eric Jorgenson's Life of Leverage course about? Life of Leverage is a course by Eric Jorgenson, author of the Navalmanack, that extends Naval Ravikant's four leverage types (people, capital, code, and media) into a practical framework for operators. It teaches how to measure whether you're actually using leverage, which metrics track it, and how to restructure your working hours to produce non-linear output. What are the four types of leverage Eric Jorgenson and Naval Ravikant describe? Naval describes four leverage types: people (teams and labor), capital (money and assets), code (software that scales at zero marginal cost), and media (content that reaches audiences without ongoing effort). Jorgenson's course goes deeper on each, with specific metrics and diagnostic questions to assess how well you're actually deploying them. How does the leverage framework apply to LinkedIn engagement strategy? Media is the most accessible form of leverage for most operators, and LinkedIn engagement is one of the highest-ROI expressions of it. A single well-placed comment on a high-traffic post reaches the author's audience, the commenter's network, and remains active long after posting. Operators who treat each engagement as a compounding asset rather than a one-time output get disproportionate returns over time. What metrics should operators track to measure leverage? Jorgenson highlights several: percentage of time spent on work only you can do, percentage of time on low-value or low-skill work, income per hour worked, income per hour not worked (what your systems earn while you're elsewhere), return on owned assets, and revenue per employee. The income-per-hour-not-worked metric is the one most operators ignore and often the most revealing. <hr> This article was first published on the <a href="https://www.getchime.co/blog/life-of-leverage-eric-jorgenson?utm_source=devto&utm_medium=republish">Chime blog</a>. For more on LinkedIn pipeline tactics, <a href="https://www.getchime.co/blog?utm_source=devto&utm_medium=republish">visit the Chime blog</a>.

Top 10 takeaways from The Agents #006

Chime — Thu, 11 Jun 2026 03:04:24 +0000

SaaStr published the back end of their go-to-market agent stack on episode 006 of The Agents: commit counts, API stacks, monthly costs, live demos. Three humans running more than 20 AI agents is the kind of claim that usually dissolves when you look at the receipts. This time the receipts are in the episode. We pulled the 10 numbers worth sitting with.

SaaStr's go-to-market agent stack runs at a fraction of equivalent headcount cost. Their AI VP of Marketing costs $257 a month, their inbound agent handled 402,000 conversations in a single event cycle, and one sales agent recovered $500K from leads no human would have worked. The through-line: deploy agents where volume exceeds human capacity, not just where cost is lower.

1. $257 a month replaced an entire BI workflow

10K, SaaStr's AI VP of Marketing, runs at $257 a month. That is roughly $3,084 a year, under 3% of one loaded marketing-analyst salary. It started in January as a dashboard connected to Marketo and Salesforce. Four months later it owns the number, runs daily forecasting, tracks every campaign in real time, and pushes the top three marketing ideas every morning.

Price every agent against a loaded salary, not against a free tier. $257 against $100K-plus in salary and overhead is a category difference, not a discount.

2. 402,000 conversations in one event cycle

Amelia AI, running on Qualified, fielded 402,000 chat interactions across 2.25 million sessions on annual.com for a single event. Three people cannot physically run 402,000 conversations. Even a strong BDR handles a few dozen meaningful conversations a day. At that volume, the efficiency framing misses the point: with three humans, 402,000 conversations isn't expensive, it's impossible.

3. 614 meetings, roughly $52M of theoretical pipeline

Amelia AI booked 614 qualified meetings for SaaStr Annual. At an ~$85K average sponsorship, that's roughly $52M of theoretical pipeline from one agent. Not all of it closed. But against headcount: a strong BDR books 10 to 15 qualified meetings a month, so 614 in a single event cycle is on the order of 3 to 5 BDR-years of booking output, compressed, with near-zero complaints on the booking experience.

4. 1,000 commits in 120 days, and lines of code told them nothing

10K has roughly 1,000 commits across four months, about 7 to 8 a day. Annie, their event site turned agent, went from 18,000 to 45,000 lines of code in two weeks -- yes, some of it is slop. The output didn't care. SaaStr is improving one internal application daily, not shipping to a million users on a fragile base. Commit velocity and code volume are irrelevant metrics when you're judging internal agent performance.

5. The slide that costs $1,400 per deal

SaaStr marks up and discounts in a controlled band. The problem is what happens at the close: a rep who smells a deal slipping moves from a planned 20% off to 25, then 30, then 34. The data says the extra discount doesn't move the close rate. On a $10K ticket, sliding from 20% to 34% off is $1,400 of pure margin per deal with no measurable lift. Across 100 deals, that's $140K gone.

A 20% discount off a marked-up price lands differently than a 20% discount off the base price -- and Amelia AI applies the right discount inside hard rules without negotiating against itself under pressure. SaaStr frames it as real-time CPQ. Guardrail the discount band and you stop subsidizing the anxiety of the close.

6. $500K recovered from leads humans never touched

A-leads don't need an agent. A million-dollar inbound gets a reply from your least attentive rep in 60 seconds. The value is in the B-leads: real signal, real score, but never worth a human's time, so they decay in the database. SaaStr pointed Artisan at exactly those and it returned an extra $500K against a cost in the low thousands a month.

Agents on the work humans structurally skip produce incremental revenue that didn't exist before. Agents on the deals humans already chase produce marginal uplift.

7. 150 accounts managed at under 90 days old, with zero Salesforce data

QB, SaaStr's AI VP of Customer Success, manages 150 accounts and it's under 90 days old with no Salesforce integration yet. The point isn't that it's feature-complete. SaaStr's pattern across agents is to ship into production early, measure output, and layer in integrations as the agent proves itself. They don't wait for the perfect data environment.

8. Three humans as the operating constraint, not the ceiling

The 3-human number is the framing device for everything else in the episode. SaaStr isn't using agents to reduce headcount from 30 to 3. They built the operation at 3 from the start, with agents covering the surface area that headcount would otherwise fill. That's a different architecture from "we replaced some people with AI." The agent stack is the org chart, not a supplement to it.

The question worth asking isn't how to use AI to do what you're already doing faster. It's what work you would simply never do at your current scale that an agent makes possible. The SaaStr numbers are the answer to that question.

The pattern recurs across other operators. The same architecture appears in our breakdown of AI agent stacks at SaaStr, Owner, and Klaviyo, and the Aurasell CEO episode maps the GTM stack implications directly.

9. Daily forecasting is what makes the agents accountable

SaaStr runs daily forecasting through 10K specifically because they need to know what each agent produced that morning, not quarterly. The agent only proves its value if someone is watching the output. Without that measurement layer, you can't tell whether the agent is performing or just running.

10. Output is the only metric that survives contact with production

Across every agent in the episode, SaaStr's evaluation framework is the same: what did it produce? Not how it was built, how many commits it took, how clean the code is, or whether the integrations are complete. The 402,000 conversations happened. The 614 meetings were booked. The $500K came in. Those numbers don't care about the architecture decisions that generated them.

Deploy agents where the work exceeds human capacity. That's where the numbers get interesting.

The agents that perform are the ones with clear output metrics, hard guardrails, and deployment in work humans structurally can't or won't do. The ones that underperform are the ones layered on top of work humans were already doing fine.

<h2> <a name="faq" href="#faq" class="anchor"> </a> FAQ </h2> What does SaaStr's full go-to-market agent stack actually cost per month? SaaStr has shared that individual agents like 10K (their AI VP of Marketing) run at $257 a month, and sales agents like Artisan cost in the low thousands per month. The full stack cost across 20+ agents hasn't been published as a single number, but the individual agent economics they've shared suggest total costs well under the equivalent headcount for those functions. How did SaaStr's AI agent book 614 meetings for one event? Amelia AI, running on Qualified, handled inbound chat across 2.25 million sessions on annual.com during the event cycle. By fielding 402,000 conversations automatically and qualifying leads in real time, it booked 614 meetings without BDR involvement. SaaStr equates this to roughly 3 to 5 BDR-years of booking output compressed into a single event cycle. What is the right way to measure ROI on a go-to-market AI agent? SaaStr's framework is to price every agent against loaded headcount, not against a free tier or a cheaper tool. A $257/month agent replacing a $100K marketing analyst role isn't cheap, it's a category difference. They also measure output directly: meetings booked, conversations handled, pipeline recovered. Commit count and lines of code are irrelevant. How do you prevent AI agents from over-discounting on sales calls? SaaStr found that reps sliding from a planned 20% discount to 34% off in a panic produced $1,400 in lost margin per deal with no measurable improvement in close rate. Their fix is to give the AI agent a hard discount band and not allow deviation. The agent applies the right discount inside set rules and doesn't negotiate against itself under pressure. <hr> This article was first published on the <a href="https://www.getchime.co/blog/top-10-takeaways-agents-006?utm_source=devto&utm_medium=republish">Chime blog</a>. For more on LinkedIn pipeline tactics, <a href="https://www.getchime.co/blog?utm_source=devto&utm_medium=republish">visit the Chime blog</a>.

AI is not coming to save your LinkedIn

Chime — Wed, 10 Jun 2026 03:01:24 +0000

Matt Gray runs Founder OS, a community that puts him in front of thousands of founders every month. He talks to people doing $75M in annual profit. He knows what the top of the distribution is focused on. And the thing he keeps coming back to is this: the founders who are losing on LinkedIn right now are the ones who outsourced the wrong thing to AI.

AI is not coming to save your LinkedIn presence because the thing LinkedIn rewards -- a specific point of view from a specific person with a specific track record -- cannot be generated in bulk. Tools can help you transcribe, distribute, and analyze. They cannot create the friction-free sense of "I know exactly who this person is" that makes someone click, comment, or reach out. Founders who treat AI as a content factory get ignored. Founders who treat it as infrastructure get compound returns.

The experiment everyone is running right now

Drop into Claude. Generate 40 LinkedIn posts in an afternoon. Schedule them under your name. Watch the engagement numbers come in flat.

We've seen this pattern across dozens of the profiles we audit. The posts are grammatically clean. The hooks follow a recognizable structure. The calls-to-action are present. And the comment sections are quiet.

The audience doesn't articulate why they're tuning out. They just scroll past. What they're responding to -- or failing to respond to -- is the absence of something specific: the founder's actual perspective, arrived at through their actual experience, in their actual voice.

LinkedIn's audience is not unsophisticated. The people you're trying to reach -- buyers, future collaborators, referral partners -- have read enough generated content to recognize its texture. It's not that AI-written posts are obviously bad. It's that they're obviously interchangeable. And interchangeable content gets skipped.

What you actually have that AI doesn't

Gray puts it plainly: "My soul is the only thing I have that nobody else can copy."

That sounds like founder-newsletter earnestness until you think about what it actually means for distribution mechanics.

Your competitive surface on LinkedIn is the intersection of three things that are genuinely hard to replicate:

Your specific vantage point. Not "founder" in the abstract, but the particular set of decisions you've made, customers you've served, and bets you've lost. Gray talked to Neil Patel this week and learned something specific about how Patel has pulled $75M in profit annually for 16 consecutive years. That fact, in Gray's newsletter, from Gray's relationship, carries weight that a generated version of "high-profit founders think differently" does not.

Your earned credibility on a narrow topic. The operators who build real inbound pipelines through LinkedIn are not generalists. They have a specific claim to authority -- a niche, a methodology, a type of customer they serve better than anyone else -- and they repeat it in different forms until their audience associates them with it automatically.

Your actual opinion, including the uncomfortable parts. Generated content optimizes for plausibility, which means it optimizes for the center of the distribution. The posts that drive comment threads -- and the profiles that generate inbound -- tend to hold positions that are specific enough to be wrong. AI tools don't produce those by default. You do.

Where AI is actually useful

Gray isn't anti-AI. He uses it daily. The distinction he draws is precise: AI helps him with infrastructure, not with the thing he's distributing.

He records voice memos on his walks. AI transcribes them. He uses Claude to analyze his best-performing YouTube content so he can understand what's working. He builds dashboards. He processes. He does not ask the tool to have his opinions for him.

This is the frame that matters for LinkedIn operators specifically. The bottleneck is not content production. Most founders can write a solid LinkedIn post in 20 minutes if they already know what they want to say. The bottleneck is knowing what to say, knowing where to say it, and knowing which conversations to enter.

That's the work AI can support without replacing the thing that makes the content worth reading.

Specifically:

Transcription as ideation capture. If you think clearly on walks, in the shower, or in client calls, voice-to-text tools mean you don't lose the thought. The idea is yours. The transcript is infrastructure.

Pattern analysis on your own content. What did your last 30 posts have in common when they worked? AI can surface that faster than manual review. But the insight only matters if you have 30 posts to analyze -- which means the human work has to come first.

Research and synthesis on topics you're already expert in. AI is faster at pulling context around a topic you already understand than it is at generating genuine expertise you don't have. Use it to supplement, not to originate.

Editing for clarity, not for voice. If a post is clear in your head but muddy on the page, AI can help untangle the structure. The trap is asking it to rewrite the post entirely, at which point you've handed over the one variable that was actually doing the work.

The mechanics of why this matters for pipeline

LinkedIn inbound works through a specific mechanism: someone in your target audience sees you in a comment section (usually on a post by someone they already follow), reads what you wrote, thinks "this person clearly knows something," and clicks through to your profile. From there, a subset of them follow you, a subset of those read your content over time, and a subset of those reach out when they have a problem you can solve.

Every link in that chain depends on the comment being genuinely good -- specific, grounded in real experience, and distinct from the other 40 comments on the post. A generated comment is rarely any of those things.

We've written about what top LinkedIn creators actually do differently and the finding that keeps showing up is the same: the accounts that generate real pipeline are doing less, not more, but the less they're doing is more specific. Fewer posts, clearer point of view, more time spent in the right comment sections.

The profiles we see generating the most inbound are not the ones with the most content. They're the ones where every post and every comment traces back to a coherent identity the reader can understand in about 10 seconds.

You cannot manufacture that identity at scale with AI. You can document it, distribute it, and make it easier to find. That's the job.

What this means practically

If you're spending time generating posts in bulk, the question worth asking is: what are you actually trying to build? If the answer is "a body of content that represents my perspective clearly enough that the right people find me and reach out," then volume is not the metric. Clarity is.

The operators we see building real inbound pipelines on LinkedIn typically share one pattern: they have a specific claim they're willing to repeat in different forms across different posts and different comment sections, and they don't dilute it by generating off-brand content to fill a schedule.

Matt Gray's newsletter gets forwarded. The posts in it don't feel like they came from a content factory. That's not an accident and it's not magic. It's the consequence of writing every word himself, even when that means writing less.

Gray's formulation is worth keeping: pick three places where AI is genuinely useful, and ignore the rest. For most LinkedIn operators, those three places are something like transcription, content analysis, and research support. Not drafting. Not voice. Not perspective.

The audience you're trying to reach has enough content. What they don't have enough of is someone they can actually trust. AI can't close that gap. You can.

For a closer look at how this plays out in practice, our audit of Justin Welsh's LinkedIn strategy shows what a clear, human-specific point of view looks like when it compounds over time. And if you want to see the comment-section mechanics that actually drive inbound, the founder-led brands playbook covers the specific patterns we've tracked across profiles in several niches.

FAQ

Can AI write my LinkedIn posts if I give it my past content to train on?

It can approximate your surface patterns -- sentence length, topic areas, common phrases -- but it can't replicate the thing that makes your best posts work: the specific experience or opinion that only you arrived at. Training AI on your past content produces content that sounds like you, not content that is you. The distinction matters because your audience responds to the latter.

How do I know if my LinkedIn content is too AI-generic?

Read your last five posts and ask: could any other founder in your niche have plausibly written this exact sentence? If the answer is yes for more than one sentence per post, the content is too generic. Strong LinkedIn posts have at least one claim or example that only you could have produced -- a specific client situation, a specific number, a specific opinion that someone could reasonably disagree with.

Where should I actually use AI in my LinkedIn workflow?

Three places work well: transcribing voice memos so you don't lose ideas from calls or walks, analyzing your own past content to find what's working, and doing background research on topics you're already an expert in. Where it tends to hurt: drafting posts from scratch, writing comments on other people's content, and generating 'variations' of a post that dilutes the original point.

Is posting less on LinkedIn actually a better strategy for inbound?

Across the profiles we audit, the relationship between post volume and inbound quality is weak. What predicts inbound is clarity of point of view and consistency of that point of view across posts and comments over time. A founder posting three times a week with a sharp, specific perspective typically outperforms one posting daily with generic content. The schedule matters less than the signal.

This article was first published on the Chime blog. For more on LinkedIn pipeline tactics, visit the Chime blog.

SaaStr built an AI VP of marketing

Chime — Wed, 10 Jun 2026 03:01:23 +0000

SaaStr spent over $500K on AI infrastructure before concluding that no off-the-shelf marketing agent could do what they actually needed. So they built one from scratch and named it 10K. What they found is worth paying attention to if you're building inbound pipeline right now.

SaaStr's AI VP of marketing, called 10K, handles daily campaign planning, pipeline prediction, and cross-channel orchestration using five-plus years of historical data fed into Claude Opus and a Replit app. It is not primarily a content generator. Its job is strategy and execution sequencing, not copy. It outperforms an average human marketer on data processing and agenda-free analysis, but runs alongside a human VP rather than replacing one.

The content factory mistake

SaaStr has 5,000-plus pieces of content across 13 years. Twenty million words. When they surveyed the AI marketing agent market, every tool they demoed did the same thing: write more content. Blog posts, social captions, email copy, some SEO optimization.

The problem is that content volume was never their bottleneck. It's rarely anyone's real bottleneck once they've been at this for more than a year.

Content generation is the easiest part of the problem to automate. It's also the least valuable part to automate. AI marketing tools defaulted to content because what buyers are easiest to impress with in a demo is a polished blog post in 30 seconds. A tool that tells you your Tuesday email campaign is cannibalizing your Friday LinkedIn window is harder to show in 20 minutes. So the industry optimized for the former.

SaaStr's 90/10 rule is worth stealing: buy 90% of what you need, build only the 10% where no off-the-shelf solution exists. They bought aggressively and found the 10% the market hadn't solved. That gap, between content creation and marketing orchestration, is where they built.

What 10K actually does

10K runs on Claude Opus analyzing five-plus years of SaaStr's historical data, fed into a Replit app. Every day it:

Plans the week's and quarter's marketing activities
Builds campaigns with specific offers, messaging, and targets
Assigns specific tasks to specific humans with dates attached
Connects to Salesforce and predicts which sponsors are likely to convert and when
Updates in real time as new pipeline and vendor data comes in

The last point matters more than it looks. Most marketing plans are built monthly or quarterly and then defended against reality rather than updated by it. 10K updates daily. When a campaign underperforms, it flags it on day two, not day sixty.

That distinction is where most human marketing operations break down. There's a natural bias toward defending the plan you made, especially when someone on the team has been running a specific channel for three years. 10K has no sunk cost to protect. It reads the pipeline data and tells you which channel is outperforming by 4x without caring that someone built their identity around the other one.

Zero agenda is genuinely valuable. Most marketing teams don't have access to it and wouldn't easily admit they're missing it.

Is it better than a human VP?

Not exactly, and SaaStr is clear about this. Better than an average human marketer on data processing and unbiased analysis, yes. Better than a great VP of marketing, not yet.

The more interesting framing is that the question itself is wrong. SaaStr runs 10K in parallel with a human VP. Both produce recommendations. Both get compared. The pitch from most AI vendors is that AI replaces expensive headcount. At SaaStr, AI makes the human sharper by providing a second opinion with no political stake in surviving the next review cycle.

For operators building inbound pipeline, the useful translation starts here: the goal is not to generate more content. The goal is to show up in the right places at the right time with enough consistency that the right people start recognizing you. That requires sequencing intelligence, not just output volume. Understanding how LinkedIn inbound signals work is the foundation before any automation layer makes sense.

Where it falls short

SaaStr is honest about the gaps. There's no central dashboard yet for managing all AI agents across the business. 10K handles marketing data well; it doesn't yet integrate cleanly with every other agent running in parallel.

The subtler gap is data quality. 10K is as good as the data it has. Five years of SaaStr's historical data is a genuinely rich training set. Most operators don't have that. If you've been posting on LinkedIn for eight months with inconsistent formats and no systematic tracking of what landed, an AI orchestrator has thin material to work with.

This is why auditing your existing presence before adding any automation layer is the right order of operations. The operators who get the most out of systematic engagement have usually done the work of understanding their current pattern first: what's working, what's been buried, which comment threads generated actual conversations versus isolated likes. A quick bottleneck audit is the precondition for useful orchestration.

Every model improvement makes 10K more precise; every week of new data sharpens its predictions. The compounding is the point. The asset being built is the data layer, and that asset compounds in a way that content generation alone never does.

What operators should take from this

The market for AI marketing tools is optimized for demo-ability, not for the actual bottleneck. The easiest tools to buy are solving the wrong problem for most operators who have been at this for more than a year.

Agenda-free analysis is genuinely scarce. Anything you can use to get an honest read on what's working is worth the friction.

The data layer compounds. The reason to start tracking your engagement behavior systematically is not to optimize this week's comments. It's because six months of clean data is worth more than six months of scattered activity when you want to understand what's actually driving inbound. The patterns that drive LinkedIn inbound become visible once you have enough signal. Before that, you're guessing.

SaaStr's 10K is a specific solution to a specific problem. But the underlying logic applies broadly: stop automating the wrong thing, build the data asset, run AI recommendations alongside human judgment rather than replacing it.

The market optimized AI marketing tools for demo-ability, not for the actual bottleneck. The easiest tools to buy are solving the wrong problem.

FAQ

What is an AI VP of marketing and what does it actually do?

An AI VP of marketing is a system that handles marketing orchestration: campaign planning, pipeline prediction, channel sequencing, and daily task assignment, rather than just content generation. SaaStr's version, called 10K, analyzes years of historical data using Claude Opus and updates its recommendations daily based on live pipeline and vendor inputs. It connects to Salesforce, predicts which prospects are ready to engage, and tells the human team exactly what to do each day. The distinction from content-generation AI tools is that orchestration is the job, not copy production.

Can an AI replace a human VP of marketing?

Not entirely, and the more useful framing is probably not the goal. SaaStr runs their AI system in parallel with a human VP, comparing outputs and debating recommendations. The AI outperforms on data processing and unbiased analysis because it has no political stake in any particular outcome. A human VP brings relationship context, situational judgment, and accountability that a model can't replicate. The combination, not the replacement, is where the real value is right now.

Why do most AI marketing tools just generate content instead of handling strategy?

Content generation is easy to demonstrate in a sales demo. A tool that writes a blog post in 30 seconds is immediately legible to a buyer. A tool that tells you your Tuesday email campaign is cannibalizing your Friday LinkedIn window requires the buyer to already understand their own data well enough to evaluate the recommendation. The market optimized for demo-ability, which skewed product development toward the easiest-to-show feature rather than the hardest-to-solve problem. Orchestration, knowing what to do when across which channels, is where the real gap remains.

How should solo operators or small teams think about AI marketing orchestration?

Start with the data layer before adding any orchestration. An AI system is only as useful as the signal it has to work with. If your engagement history is scattered and untracked, any tool trying to make predictions from it will produce thin recommendations. The right order is: audit your current pattern, understand what's actually generating inbound versus what's generating likes, then introduce tooling that can act on that signal. For LinkedIn specifically, that means knowing which comment threads led to real conversations, which post formats reached your buyer's network, and whether your engagement cadence is consistent enough to build recognition.

This article was first published on the Chime blog. For more on LinkedIn pipeline tactics, visit the Chime blog.

$400M ARR, under 200 people: Elena Verna's B2B playbook

Chime — Tue, 09 Jun 2026 03:03:43 +0000

Elena Verna spoke at SaaStr AI a year into her role as Head of Growth at Lovable, right around when the company crossed $400M ARR with fewer than 200 people. The question she put in front of the room: when AI writes 80%+ of your code and anyone can vibe-code your feature set over a weekend, what is actually left to compete on?

Lovable hit $400M ARR with under 200 people by rejecting feature differentiation as a durable moat and leaning hard on brand, data, network effects, and org structure that traditional headcount ratios can't replicate. Head of Growth Elena Verna argues that the companies still building around feature leads are building on borrowed time, and that the next decade of B2B belongs to orgs that ship fast, stay lean, and invest in the moats AI can't clone.

The $2M ARR per employee number is not a coincidence

The math is simple and the implications aren't. Lovable had north of $2M in ARR per employee at $400M. That ratio wasn't possible in the pre-AI org chart, and it didn't happen by accident.

Verna calls their structure "product engineering." The old ratio, one PM to seven engineers to a designer to a marketer, is gone. Everyone does IC work. Everyone ships. No one is purely coordinating.

What makes this stick is that the model only works if the people at the top stay builders. Verna is explicit: she fired herself out of the marketing job, then handed off the growth lead role, and went back to being an individual contributor. That's not a demotion story. It's a deliberate bet that her leverage is higher as a high-powered IC with a stack of agents than as someone running coordination meetings.

The revenue-per-head number is the output of that bet, not the goal. When everyone ships to production, the metric takes care of itself.

Feature differentiation is a short-term lead, not a moat

For the past 15 years, B2B companies built growth engines on top of product advantages. We won because we had better engineers, a better roadmap, faster execution on features. Verna's argument is that this worked because building was expensive and slow. Neither is true anymore.

When 80%+ of your code is written by AI, a feature lead is worth weeks, maybe a few months. Then a competitor with the same AI access closes the gap. You can still get ahead on features. You just cannot build a predictable, compounding growth engine on top of something that erases in a sprint.

The moats Verna says still hold:

Hardware. Still genuinely difficult and capital-intensive to build.
Network effects. Always hard to create. Once real, they compound on their own.
Proprietary data. The kind competitors cannot replicate regardless of how good their models get.
Security and compliance. Slow, expensive, and worth it for exactly that reason. The switching cost is real.
Brand. "Brand is back, baby" is how Verna put it. When everyone can build the same product, the relationship with the customer is what's left.

Notice what's not on that list. SEO and SEM are absent, and Verna is deliberate about that. When AI-generated content is everywhere and AI is eating the top of the search funnel, channel strategies built on organic and paid search are more fragile than they looked three years ago.

For operators building on LinkedIn, that absence is worth noting. Earned authority is how you build brand without a media budget, and it doesn't evaporate when a competitor matches your feature set.

The flat org is not a culture gimmick

Lovable runs with no internal titles. The reason is structural, not philosophical. When everyone is expected to actually build, titles become noise. The signal is what shipped.

A few mechanics Verna describes that make the velocity real:

A #shipped Slack channel where the day's production releases pile up. Multiple ships a day, not a sprint cadence.
No separate layers of approval between an idea and production. The IC ships.
Leaders who want to stay in leadership need to stay in the work. Coordination without building is not a long-term job at Lovable.

The org design matters for operators outside Lovable too, because it names a problem a lot of teams are quietly carrying. The traditional reward for being a great individual contributor has always been a promotion into management. Management is a different job. Most people were not built for it, and a lot of them know it.

Verna's advice to founders looking for this kind of person: go find the leaders who are quietly miserable in coordination roles and ask if they want to build again. She says a surprising number are ready to say yes, because they see it as a way to fall back in love with the work.

PMF is not a finish line

Even approaching half a billion in revenue, Verna says Lovable is still "on the product market fit treadmill." The category is moving fast enough on both the technology and customer sides that they feel like they have to recapture PMF every month.

This reframes something a lot of growth operators get wrong. PMF is treated as a milestone you cross and then leave behind. You raise the Series B, you scale the team, you shift to execution mode. Verna's point is that in AI-native categories, scale doesn't let you slow down. It raises the stakes on velocity, because the window between "we have PMF" and "the category has moved" is shorter than it used to be.

The same logic applies to content strategy. The signal in comments and engagement tells you whether your angle still fits the moment, and ignoring it is how you get caught flat when the category shifts.

What this means if you're not Lovable

The $400M ARR number is real and the org is genuinely unusual. But most of what Verna is saying translates directly to founders and operators at much smaller scale.

The moats that hold (brand, network, proprietary data) are available to a solo operator. A founder with a strong LinkedIn presence has brand. An operator who shows up consistently in the right comment sections has a network effect in their niche. A consultant who publishes observations from actual client work has proprietary data that a competitor with the same AI access cannot replicate.

The org point translates too. The operators with the highest LinkedIn conversion tend to do the IC work themselves, writing their own comments, building their own lists, staying close to what's working. Delegation without a feedback loop produces coordination overhead, not velocity, regardless of team size.

For a worked example of what lean, high-conviction LinkedIn presence looks like in practice, the breakdown of Kyle Poyar's 107K strategy covers the mechanics in detail. The founder-led brands LinkedIn inbound piece is the complementary frame on why the approach compounds.

Verna's talk is the macro argument. The question for most operators is what it changes about what you do Monday morning. Stop betting on features as your differentiation story, start building the things AI can't clone, and ship something before the week is out.

FAQ

How did Lovable reach $400M ARR with under 200 people?

Lovable achieved more than $2M in ARR per employee by rejecting traditional org structures. Every team member does IC work and ships to production directly. There are no internal titles and no layers of approval separating an idea from a production release. The model only works because leaders, including Head of Growth Elena Verna, stay builders rather than pure coordinators.

What moats does Elena Verna say still work in B2B when AI can clone features quickly?

Verna names five moats that hold even when AI collapses the cost of building: hardware (still capital-intensive), network effects (hard to create, compounds once real), proprietary data (competitors can't replicate it regardless of their models), security and compliance (slow and expensive, which is the point), and brand (when everyone can build the same product, the customer relationship is what remains). SEO and SEM are explicitly absent from her list.

What is the 'product market fit treadmill' Elena Verna describes?

Verna uses the term to describe Lovable's experience of having to recapture PMF continuously rather than treating it as a milestone they passed. Even at $400M ARR, the category moves fast enough on both technology and customer behavior that the fit they had last month may not hold next month. It reframes PMF as an ongoing calibration rather than a one-time threshold.

What does Lovable's flat org and no-titles structure mean for B2B operators?

Lovable's no-title structure is a structural choice, not a culture statement. When everyone is expected to ship, titles are noise. For operators outside Lovable, the principle is the same: the highest-leverage version of any role is one where the person doing it stays close to the actual work. Verna's broader point is that the next decade's career flex is the high-powered IC with agents, not the VP title earned through coordination.

This article was first published on the Chime blog. For more on LinkedIn pipeline tactics, visit the Chime blog.

Anthropic and OpenAI are hiring GTM roles

Chime — Tue, 09 Jun 2026 03:03:42 +0000

We pulled open-role data from Anthropic and OpenAI's current job boards to see where each company is concentrating its hiring. The answer runs counter to everything the AI-replaces-sales narrative assumes.

Anthropic and OpenAI are each hiring go-to-market roles faster than any other department. At OpenAI, roughly one in five open roles sits across sales, partnerships, and revenue functions. At Anthropic, sales accounts for approximately 20% of all open positions. These are the two strongest product organizations in AI, and their hiring pattern confirms that strong product does not solve the distribution problem.

The hiring data

Anthropic shows the pattern clearly. Sales represents roughly 20% of all open roles, more than engineering, safety research, or any other department. For a company that built its public identity around responsible AI and frontier research, the commercial staffing mix is a deliberate shift in emphasis.

OpenAI is reportedly planning to nearly double its headcount this year, growing from roughly 4,500 employees toward 8,000. A significant portion of that growth is customer-facing, with enterprise account executives among the most aggressively recruited positions. A company walking into public markets at that scale needs revenue infrastructure, not just research credentials.

On June 1, Anthropic filed confidentially for a U.S. IPO, days after a $65 billion funding round pushed its valuation toward $965 billion. Both companies are building the full enterprise stack alongside active frontier research: account executives, partnerships, revenue operations, customer success, and field marketing.

What this means for operators

Strong product does not solve weak distribution. These two companies have the most defensible AI products on the market and still need a full GTM team to move enterprise deals.

That logic holds at every scale. The operators we see generating the most inbound from LinkedIn are not the ones with the strongest credentials or the most sophisticated offering. They are the ones who have invested in their distribution surface consistently, while their competitors assume the work will speak for itself.

If Anthropic, with $65 billion in fresh funding and a genuinely state-of-the-art model, still needs 20% of its hiring in sales, no one is exempt from the distribution problem.

LinkedIn is where most enterprise buyers are spending professional attention. The operators showing up consistently in the right comment sections, building relationships with the right audiences, and generating inbound from that activity are building something that compounds. We have written about LinkedIn inbound signals and what separates profiles generating conversations from those generating only impressions. The pattern holds across the accounts we audit: distribution is deliberate work, not a byproduct of being good at your job.

Anthropic and OpenAI are allocating headcount to distribution at the same rate they allocate it to research. The question for operators is whether they are applying the same logic to their own time. A single founder building pipeline through LinkedIn faces the same mechanic. Distribution compounds only if it is treated as deliberate work.

For context on what that looks like at the individual level, the founder-led brands LinkedIn inbound patterns we have documented show the same dynamic: the accounts generating pipeline are the ones that treat distribution as a core function, not an afterthought.

FAQ

Why are Anthropic and OpenAI hiring so many GTM roles if AI is supposed to automate sales?

Building AI that can assist in sales workflows is a different problem from actually distributing an enterprise product. Both Anthropic and OpenAI are selling to large organizations that require account executives, deal management, and relationship infrastructure regardless of how sophisticated the underlying technology is. Their hiring data shows that strong product quality does not solve the distribution problem automatically.

What percentage of OpenAI's open roles are in sales and GTM?

Roughly one in five open roles at OpenAI sits across sales, partnerships, and revenue functions, making go-to-market the largest single hiring category at the company. Enterprise account executives are among the most aggressively recruited positions as the company moves toward doubling its total headcount this year.

What does Anthropic's IPO filing tell us about its commercial strategy?

Anthropic filed confidentially for a U.S. IPO on June 1, days after a $65 billion funding round pushed its valuation to approximately $965 billion. Sales now accounts for roughly 20% of all open roles, more than any other department, suggesting the company is building the revenue infrastructure needed to justify its public-market valuation.

What can founders and operators learn from how Anthropic and OpenAI are hiring?

The hiring pattern confirms that distribution is a separate problem from product quality, and that even the strongest products require active investment in GTM to reach buyers. For operators building pipeline through LinkedIn and content, the same principle applies: expertise does not distribute itself, and the accounts generating consistent inbound are the ones treating distribution as a core function rather than an afterthought.

This article was first published on the Chime blog. For more on LinkedIn pipeline tactics, visit the Chime blog.