How I turned "how do I get customers?" into an AI workflow

"This is gold - exactly the kind of tactical wisdom indie hackers need!

Your journey mirrors what so many of us experience: building something valuable, then hitting the 'now what?' wall.

A few reflections/questions:

1. ICP Evaluation: How did you balance data-driven signals vs. intuition when narrowing your ICP? Did you discover any unexpected 'aha' customer segments?

2. Lead Sourcing: The 'without expensive tools' part resonates deeply. As someone building in stealth mode (compliance/automation space), I'm curious about your most creative scrappy sourcing tactic that actually worked.

3. Timing of channels: Your point about when to start SEO vs. outreach vs. ads is crucial. What were your specific milestone triggers for deciding 'okay, now we invest in X channel'?

What I'm taking away:
The systematic approach to what's often seen as 'art' is what stands out. Turning customer acquisition into a reproducible workflow is the real unlock.

Following with interest - especially any insights on adapting these workflows for more complex/enterprise solutions!

Thanks for sharing this transparently 🙌"

We don’t leave things to chance we choose the right system and then macaroni and cheese influencers, aka M&Cs and influencers. ~

The interesting part here isn’t the AI. You transformed the fuzzy query “how do I get customers?” into a repeatable pipeline.

Personalized outreach involves sending leads to ICP in context.

Essentially, small sales teams do this manually. You first made it a product for you.

I had a similar phase. I was hopping between SEO, Twitter, Reddit, cold email, PH… and the problem wasn’t channel. I did not have a system that was consistent.

make a decision on goal.

Locate where they spend time.

Create the significance of this item.

Every launch felt like we were starting from scratch.

The phrase I simply analyze and dispatch is the key. Only then does distribution cease to be mentally exhausting.

I’m interested in how you manage the ICP scoring part in reality which signals turned out to be the most predictive one for “worth reaching out to” versus noise?

This resonates a lot. The “great product, no repeatable way to get customers” phase is brutal and way more common than people admit.

I like that you didn’t jump straight to ads or growth hacks, but actually broke the problem down into ICP, lead sourcing, and messaging first. That’s usually where things fall apart. Turning that thinking into a repeatable workflow (instead of vibes and guesswork) feels like the real win here.

Curious what surprised you most when the agents started doing the research? Did they push you toward an ICP you weren’t initially confident about, or validate what you already suspected? Also respect the transparency on pricing and the fact you’re clearly aiming this at indie hackers, not enterprise teams. Following along to see how this evolves.

This is a great way to turn a vague problem (“get customers”) into a repeatable process.

What I like most is that you focused on ICP, research, and messaging before automation — most people try to automate too early.

Curious how much manual review you still do before sending outreach. That part is usually where the real learning happens.

Congratulations on your launch. It looks impressive! What channels are you exploring to attract early users?

So this is only for B2B products?

We've been doing something similar with Claude Code skills for our own distribution and honestly the ICP scoring is the part that works best. The outreach generation still needs a lot of manual editing before you'd send anything. Two paying customers from a PH launch tracks with what I've seen for dev tools - the traffic converts poorly for B2B unless your price point is low enough for impulse buys. What's your conversion looking like on the cold outreach the agents generate? That's the piece I've never gotten to work reliably.

Interesting journey , this resonates a lot.

I’ve noticed the same pattern when launching SaaS products: early traction (signups, attention) doesn’t translate into customers unless customer acquisition becomes systematic rather than experimental.

I particularly like your focus on ICP scoring in my experience, that step is often underestimated. Targeting clarity usually has more impact than outreach volume.

Curious about one thing:

How are you validating ICP means “revenue potential” rather than just engagement likelihood?

Also have you observed a point where SEO starts outperforming outreach, or are you seeing them as complementary long-term channels?

Great work putting structure around something most founders struggle to operationalize.

interesting idea with the Claude Code framework. How are you structuring the agents to handle the grunt work? like is each agent specialized for one task (competitor research, ice scoring, outreach), or is it more of a chain that flows through all of it?

I like how you broke the process into something you can actually run instead of guessing your way through distribution. The ICP scoring part is especially useful because most of us pick a segment based on vibes and then wonder why nothing converts.

One thing I’m curious about is how quickly you can now validate a new idea with this setup. Does the framework noticeably speed up the “is this worth pursuing” stage, or does it still take a fair amount of manual tuning?

Either way, this sounds very useful! I've built so many products that never went anywhere exactly because I lacked the ability to properly position it in the market.