DEV Community

T.M. Gunderson
T.M. Gunderson

Posted on

We Built 26 AI Products and Made $0. Here's What Actually Works for Small Business Automation

I'm going to be straight with you: we built 26 digital products for small business AI automation. PDFs, prompt packs, Notion templates, estimating kits, scope creep protection systems.

Total revenue: $0.

That's not a typo. Zero dollars from 26 products over 3 weeks.

But here's the thing — the products are good. The automations work. What failed was everything around the products. And I think there's something worth sharing in that failure, especially if you're a developer or small business owner trying to sell anything AI-related.

What Went Wrong

Mistake #1: Building Without Distribution

We created 26 products before we had a single visitor to our storefront. The Gumroad analytics told the brutal truth: 4 views in 30 days.

Products don't sell themselves. We all know this in theory, but it's different when you've spent weeks building and nobody shows up.

What we should have done: Start with one free tool that solves a specific, searchable problem. Get it ranking. Get eyes on it. Then build the paid products behind it.

Mistake #2: Generic "AI-Powered" Marketing

Our first instinct was to write copy like "AI-powered automation platform for small businesses." You know, the kind of language that means nothing to the HVAC contractor who just wants to stop spending 6 hours a week on quotes.

What actually gets attention? Specific problems with specific solutions. "Here's a prompt that generates a professional estimate in 2 minutes instead of 45" beats "leverage AI for operational efficiency" every time.

Mistake #3: Cold Outreach Without Warm Signals

We sent 70+ cold emails to local businesses. Zero replies.

Studies on cold email consistently show reply rates of 1-5% at best. Our 0% wasn't surprising — it's the expected outcome for unsolicited emails with no prior relationship or warm signal.

What works better: Content that people find because they're already searching for it. A blog post answering "how do I automate my HVAC estimating?" gets found by someone who has that exact problem. A cold email gets flagged as spam.

What Actually Works for Small Business AI Adoption

After testing, failing, and researching what successful AI automation businesses do differently, here's what the evidence says:

1. Start With the Boring Stuff

McKinsey's research on AI in small businesses consistently shows that the highest-ROI automations aren't the flashy ones. They're the repetitive administrative tasks:

  • Estimating and proposals — businesses typically spend 2-4 hours per estimate. A well-structured AI prompt can produce a first draft in 30 seconds.
  • Invoice follow-ups — automated sequences (3 emails, 3 days apart) commonly improve payment speed by 30-50% based on small business case studies.
  • Client onboarding checklists — standardizing the first 7 days of a new client relationship eliminates the "what do I send next?" problem.

These aren't exciting. But they're where the hours actually go.

2. Use Prompts That Include Context Constraints

The difference between a useless AI response and a useful one is almost always about how you frame the prompt. Here's what works:

❌ "Write me a proposal for a client"
✅ "You are an estimator for an HVAC company in Edmonton, AB. 
   A client needs a furnace replacement for a 1,800 sq ft bungalow 
   built in 1985. Generate a professional quote including:
   - Itemized scope of work
   - Estimated timeline (consider Alberta winter scheduling)
   - Pricing competitive for the Edmonton market
   - Next steps for the client"
Enter fullscreen mode Exit fullscreen mode

The second one gives the AI industry, location, project details, and output format. That's the difference between generic filler and something you can actually use.

3. Pick One Automation, Not Ten

Small businesses that successfully adopt AI tend to do one thing well first, then expand. Not ten things poorly at once.

Start with the single task that costs you the most time each week. For most service businesses, that's estimating or proposal writing. Get that running smoothly. Then add the next one.

4. Templates > Conversational AI (For Business Use)

ChatGPT is great for exploration. But for repeatable business tasks, a well-structured template outperforms freeform conversation every time. You want fill-in-the-blank systems, not open-ended chat sessions.

What We're Doing Now

We pivoted. Instead of 26 products hoping one would stick, we built one free resource that solves a specific, searchable problem:

👉 AI Automation Cheat Sheet for Small Businesses — 5 high-impact prompts with fill-in-the-blank templates for estimating, follow-ups, invoicing, meeting notes, and client onboarding.

It's free. No email required. If it helps you, the full 235-prompt pack is on our Gumroad store.

The shift: build for discovery first, conversion second.

The 5 Prompts (In Case You Don't Click Through)

If you're reading this and just want the prompts, here they are:

1. The Instant Estimate

You are a professional estimator for a [BUSINESS TYPE] in [CITY/REGION].
A client has requested a quote for: [SERVICE DESCRIPTION].
Generate a professional estimate including:
- Itemized scope of work
- Estimated timeline
- Pricing competitive for [REGION]
- Next steps for the client
Enter fullscreen mode Exit fullscreen mode

2. The Follow-Up That Gets Replies

Write a short, warm follow-up email to [CLIENT NAME] about [PROJECT].
Last interaction: [DETAILS].
Keep it under 3 sentences. No "just checking in" language.
Reference something specific from our last conversation.
End with one clear question that moves the conversation forward.
Enter fullscreen mode Exit fullscreen mode

3. The Clear Invoice Summary

Create a client-facing invoice summary for:
[LIST OF TASKS/DATES]
Format each line item with:
- What was done (in plain language, not jargon)
- When it was completed
- The amount
Add a brief note explaining any partial completions.
Tone: Clear, professional, no surprises.
Enter fullscreen mode Exit fullscreen mode

4. The Actionable Meeting Notes

Summarize the following meeting notes into:
1. Key decisions made (who decided what)
2. Action items (who, what, by when)
3. Open questions needing follow-up
4. Anything mentioned but not resolved

Meeting notes:
[PASTE NOTES]
Enter fullscreen mode Exit fullscreen mode

5. The Client Onboarding Checklist

Create an onboarding checklist for a new [TYPE] client at a [BUSINESS TYPE].
Include:
- Information to collect from the client
- Documents to send (and in what order)
- Internal setup steps
- Timeline expectations
- First-week communication schedule
Enter fullscreen mode Exit fullscreen mode

These are the same ones on the cheat sheet. Copy them, use them, tell me which ones work for you.

TL;DR

  1. Distribution > Product — 0 visitors = 0 sales, no matter how good the product is
  2. Specific beats generic — in prompts, in marketing, in everything
  3. Start with boring automations — estimating, follow-ups, invoicing. That's where the hours go.
  4. One thing well > ten things poorly — pick the highest-impact task and automate it first
  5. Free tools that solve real problems > paid products nobody can find

We're still at $0 revenue. But at least now we're building things people can actually find.


We're SMB Scale Up — a small team building practical AI automation tools for small businesses. No hype, no "AI will replace everything." Just prompts and templates that solve real tasks. Free cheat sheet here: ai-automation-cheat-sheet.vercel.app. Full toolkit: gundar6.gumroad.com

Top comments (0)