I run GTM for an AI-first company that sells to small and mid-size businesses. Which means most weeks I'm on calls with owners who want to know what AI they should be buying.
I've started giving everyone the same answer:
"Don't buy AI yet. Buy lead response time first. The AI can wait."
This usually surprises them. They came expecting a pitch. They're getting told to spend $300 a month on something boring instead.
Here's why I've landed on this answer.
## The math is brutal and well-known
The speed-to-lead research is some of the most reliable data in B2B sales. The numbers have been replicated for two decades:
- A lead reached in 5 minutes or less converts at roughly 35%.
- A lead reached in 30 minutes converts at about 21%.
- A lead reached in 1 hour converts at about 14%.
- A lead reached in 24 hours converts at about 3%.
The curve falls off a cliff between minute 5 and minute 60.
If you're an SMB with 15 inbound web inquiries a week at a $1,400 average ticket and a 30% close rate when you reach a lead in time, the difference between "answer in 5 minutes" and "answer in 24 hours" is roughly $30,000 a year in
revenue. That's not a forecast. That's a comparison of two response-time regimes you can A/B test in 30 days.
Most SMBs are running the 24-hour version without realizing it. They have a CRM. They have a website form. The form sends an email. The email sits in someone's inbox until someone gets to it.
The form is doing its job. The follow-up is the gap.
## What I see in the data
Formation Labs ran an 8-month private beta with 187 SMB owners across HVAC, plumbing, real estate, dental, marketing agency, ecommerce, and legal. We ran a four-signal audit on each: chat widget present, lead response time, review
velocity, social cadence.
Every single one of them had a lead-response gap larger than 60 minutes.
Not most of them. All of them.
Some of them had chat widgets. Some of them had nice landing pages. Some of them spent $4,000 a month on Google Ads, which is to say they were paying Google to send them leads they then took 18 hours to call back.
The pattern was so consistent that I stopped asking owners whether they had a lead-response problem. I started asking them whether they wanted to see the data on their own business. The data was always the answer.
## Why owners chase AI instead
The honest answer is dopamine. AI feels like the future. Lead-response feels like the past. One is a project you tell people about at dinner. The other is a Tuesday-afternoon chore.
But there's also a market reason. AI consultants have a sales motion. They show up at your office with a deck. Lead-response vendors mostly don't, because the product is $300 a month and nobody can afford a sales rep to sell it. So the
loud thing in your inbox is "let's talk about your AI strategy" instead of "let's measure your response time."
The result is a generation of SMB owners who spent $20-50k on AI roadmaps before they tightened the cheapest, highest-ROI lever they had. I have watched this happen four times in the last 18 months at close range.
## The hierarchy of SMB automation spend
If you imagine a stack of automation investments ordered by ROI per dollar, the order should look something like this:
- Lead-response automation — answer inbound within 5 minutes, 24/7, even from your phone. $200-500/month.
- Review-request automation — send a Google review ask after every closed job. $50-150/month or included in your CRM.
- Appointment-booking automation — let prospects self-book a slot instead of phone tag. $0-50/month.
- Email nurture for unbooked leads — 4-7 touches over 21 days for prospects who didn't convert. $100-200/month.
- Social posting cadence — keep your business looking alive online. $50-300/month.
- CRM hygiene + lead scoring — get the right leads in front of the right people. $100-400/month.
- AI agents for non-trivial customer service — handle the long tail of inbound questions without a human. $300-1,000/month.
- Custom AI workflows — anything bespoke, anything that requires consulting. $5,000+ engagement.
Most SMBs are getting pitched 7 and 8 by AI consultants. They should be buying 1 first. The ROI ratio between 1 and 8 isn't 2x. It's more like 20x in the first year.
## What "buying lead response time" actually means
You're looking for a system that does four things:
- Sees every inbound — form submissions, website chats, missed calls, sometimes email and DMs.
- Answers within 5 minutes — bot or human or hybrid. Doesn't matter as long as the prospect gets a human-quality response.
- Books an appointment when it can — the goal isn't to "engage." The goal is a calendar event.
- Hands off cleanly to a human when the conversation goes off-script or the prospect asks for one.
You're not looking for a chatbot. You're looking for a response layer that treats every inbound as a five-minute problem instead of a Tuesday-morning problem.
## Where this gets you to AI eventually
Once your lead-response layer is running, the rest of the stack gets easier to evaluate. You know what a working automation feels like. You know what an integration that actually fires looks like. You have data flowing into your CRM that
wasn't there before.
That's the right moment to start adding agents on top — review requests, email nurture, customer service. The order matters because each layer compounds on the data from the layer below.
A practical example of this stack assembled today: Sentie deploys five core agents — lead intake, social posting, review requests, email nurture, customer service — for $199-499 a month, with a human Success Manager
monitoring every account. The five-agent design isn't arbitrary; lead intake is agent number one because it's the load-bearing leg of the stack. I help run GTM for it.
But you don't have to use Sentie to follow the principle. The principle is: spend the first $300 a month on lead response, prove it works, then layer the rest.
## Closer
This is boring advice. There's no founder story, no breakthrough model, no architectural diagram.
It's also the advice that, applied seriously, would put more money in the average SMB's pocket than the last three years of AI consulting collectively did.
If your business has inbound web inquiries and an after-hours response gap, you have a math problem. Measure it for one week. Compare what you spent on Google Ads to what you closed from those inquiries. The delta is what a lead-response
system would have caught.
That's where the budget belongs. The AI can wait.
— Jim Loperfido, GTM Director, Sentie
Top comments (0)