I built a B2B AI tool. Got 500 signups in 30 days, 0 paying customers. Here's what I learned.
Three months ago, I shipped my AI tool for B2B lead generation. I posted on Twitter, Hacker News, and a few Slack communities. The growth charts looked exciting—500 signups in 30 days.
Then I looked closer. 500 signups. 3 paid. A 0.6% conversion rate that would make any marketer wince.
What I Thought Was Working
I was chasing vanity metrics. Signups felt like validation. Each new user felt like a small win.
But here's the thing nobody tells you: in B2B, a free signup isn't a warm lead. It's a "I'll try this when I have time" signal.
What Actually Went Wrong
1. I built for everyone, not someone
My tool did "lead generation for B2B companies." That's not a niche. That's a market segment the size of the ocean.
2. I didn't talk to users until week 8
I was heads-down building features. I assumed I knew what they needed. I was wrong.
3. I ignored the signup-to-paying gap
I celebrated 500 signups instead of asking: "Why aren't these people paying?"
What I'd Do Differently
The biggest lesson? Talk to 10 users before you ship anything.
I eventually rebuilt clienthunter.ai with a different approach: real conversations first, features second. Now I understand why people convert, not just why they sign up.
The Real Battle
Getting people to your tool is one thing. Getting them to pay is entirely different. I spent so much time on "growth" that I forgot to solve the actual problem—which was never the tool itself, it was understanding the buyer's psychology.
The question I should've asked from day one: "What would make you pull out your credit card right now?"
Not "what features do you want?" Not "would you use this?"—but the actual trigger for a purchase decision.
Has anyone else experienced this conversion gap in B2B SaaS? Would love to hear what's worked for you.
This is one of the most important startup lessons. Getting signups is validation for curiosity, not for willingness to pay. A lot of founders mistake attention for product-market fit.
The real turning point usually comes when founders stop asking “how do we get more users?” and start asking “what painful problem are people already desperate to solve?” At Foundersbar, we see this pattern constantly - distribution can get you traffic, but only real pain points create paying customers.
this hits — bootstrapped a b2b hosting biz for 18y and the signup-to-paid gap was always the real diagnostic, not signup count. ngl the brutal lesson for me was that "would you use this" answers basically nothing — you only learn anything once you watch someone actually try to pull out a card and stall. the "what would make you pay rn" question is the right one but tbh ppl rarely answer it cleanly — what worked better for me was reverse-engineering the 3 paid users you already have. what was diff about their context vs the other 497?
Honestly, 500 signups in 30 days for a B2B product still feels like a strong signal to me. A lot of products don’t even get people interested enough to sign up.
I think the hard part with AI tools right now is that people are curious enough to try them, but not convinced enough to pay unless the product becomes part of their workflow.
Also feels like “AI tool” alone is getting less compelling now. The products that seem to work are the ones solving one very specific painful problem really well.
I’m building something early-stage too, and I’m noticing the same thing — getting users is one challenge, getting users to change habits is a completely different one.
Talking to real users actually helps alot you will find that there is difference between the product how you want to look like and between what they really need
The "vanity metrics" lesson here is so important — and the data tells the real story if you know what to query.
500 signups vs 3 paid tells you there's a massive activation gap, but the follow-up question that most founders skip is: what did those 3 paying customers do that the other 497 didn't? That's a SQL question, not a feelings question.
The queries worth running against your user table:
When I help SaaS founders set up their data warehouse, these cohort comparison queries are usually the first thing we build — they turn "0.6% conversion" from a discouraging number into an actionable diagnosis.
The difference between a vanity metric and a meaningful one is almost always whether you have the underlying data structured to ask the next question. Happy to share a few of the diagnostic query patterns I use → free SQL scripts pack here: https://growthwithshehroz.gumroad.com/l/psmqnx
What did those 3 paying customers have in common that you noticed?
Your line about celebrating 500 signups while ignoring the signup-to-paying gap really hit. On my own tiny iOS side project — a lightweight memo app, a Captio replacement — I had a similar wake-up moment when I noticed 30+ TestFlight installs but only 3 people opened the app more than once. Free signups and free installs are basically the same illusion: they're "I'll look at this later" signals dressed up as validation. What helped me was switching from "would you use this?" to asking 5 specific users "what did you try yesterday that almost worked, and where did it break?" — those answers reshaped the next feature far more than survey-style questions. Do you think your week-8 user calls would've reshaped the actual product, or mostly the positioning?
Probably a "beginner" question, but: how did you end up "talking to 10 users" and what did you find to be the most effective approach? I've been struggling with that myself. I end up reaching out to friends or acquaintances or responding to Discord/Reddit threads, but I'm sure there's a better way...
Really appreciate you sharing this, Jack — the "built for everyone, not someone" line hits hard. It's so easy to celebrate signups until you realize they're just curious, not committed.
Your shift from "what features do you want?" to "what makes you pull out your credit card right now?" is gold. That's the real unlock.
One question for you: When you rebuilt clienthunter.ai and started having those pre-launch conversations, what was the most surprising trigger people gave you for actually paying? Like something you never would have guessed without those chats.
Jack, this hits close to home. 500 signups / 0 paid is brutal but incredibly common. The shift from "I'll try it" to "I'll pay for it" is where most B2B tools die — and it's rarely about features.
I made the same mistake (6 months building, 0 customers). That's why I built TrendyRevenue — AI validation before you write code. It checks market demand, competitor gaps (from real G2/Capterra reviews), and revenue potential in 10 seconds.
For your next feature or pivot, run it through the free tier (1 analysis, no card). It'll tell you if there's actual buyer intent, not just curiosity. The Pro plan ($39/mo) adds source-cited competitor gaps + revenue modeling — the "why" behind the signal. Might save you from another 500-signup-0-pay scenario.
The question you landed on — "what would make you pull out your credit card right now?" — is gold. Most tools stop at features. Real validation starts there.
Keep sharing the hard numbers. We need more of this.
https://trendyrevenue.com
Just shipped my first iOS app last week, a different model (B2C consumer utility), but the trap is the same. I set the paywall to fire AFTER the user finishes one successful batch conversion with their own files, not on first launch after onboarding. If they haven't felt the value yet, the paywall is just a wall.
Don't gate at signup or onboarding, gate at 'user has seen one qualified lead from this thing.' At that point, paying feels obvious.
Signups measure curiosity, paid conversion measures value acquired.
500 signups with 3 paying is not a conversion problem, it is a positioning problem. You are correctly diagnosing it.
Two things I would add from years of running and investing in early stage B2B:
The price never matched the pain. 0.6% conversion almost always means the people signing up did not feel an active fire when they hit the landing page. They felt curiosity. Curiosity is free. A burning hair on fire pain pulls the credit card. Reframe your message from "AI for B2B lead generation" (what it does) to a specific pain like "Stop wasting your sales team's first 2 weeks researching prospects who never convert." Specificity is what triggers a credit card.
The 10 user rule is right, but it has to be the right 10. Talk to people who paid for the closest existing alternative in the last 90 days, not signups. Signups are curious. Paying customers of competitors are validated. The difference between asking those two groups what they want is the difference between a roadmap and a wish list.
I see this exact conversion gap in nearly every early stage B2B AI tool I get pitched at Henson Venture. It is almost never the product. It is the lack of a sharp ICP and a pain anchored pitch.
500 signups, 0 buyers is almost always a distribution-channel problem disguised as a product problem. The signup form filtered for curiosity, not for willingness to write a check. I did the same thing last year - 400 signups from a launch post on this exact site, two paying customers, both already in my DMs before the post. The contract gets signed in conversation, never in a flow. What's the budget conversation look like on the 500 did any of them have one in a discovery call, or did the call never happen?
The signup-to-payment gap on free-trial B2B AI is brutal because the
activation moment is rarely the signup moment. Two patterns I have seen work
on non-saturated B2B audiences:
Skip the free tier entirely. Charge $49 day one with a 30-day refund
window. Counterintuitive but the "is this worth my time" question gets
answered by the price tag faster than by 14 days of trial.
Run a paid pilot offer instead of a free trial. "$500 for 3 weeks, you
get a slack channel with me, I customize the agents to your workflow, and
we both decide at the end if we continue." 1 paid pilot teaches more than
50 free signups.
The 500 free signups are valuable as a research panel. Email them a
2-question survey: "what did you expect this to do for you" and "what
stopped you from paying". Even a 5 percent response rate gives you 25
interviews worth of data.
The gap between signups and paying customers is brutal in B2B AI. I'm going through the same with Vokio — a voice AI for small business phone calls. The technical people get it immediately, but they're not the buyers. The actual buyers (shop owners, clinic managers) need to see a 30-second demo video and trust the person selling it. No amount of feature explanation converts them. Are your signups the actual decision-makers or just curious builders?
The 3 paid out of 500 detail is the important part here, because it turns this from “nobody wanted it” into “some people wanted it, but most signups were probably the wrong signal.”
I’d be careful not to treat all 500 signups as one failed cohort.
For a B2B lead gen tool, I’d want to split them by source first: Twitter, HN, Slack communities. Those channels can create very different intent. HN especially can give you people who love testing tools but have zero buying urgency.
Then I’d look at the 3 paid users and ask:
What company type were they?
What were they already using before?
What pain made them pay now, not “someday”?
Did they sign up from the same channel as the free users?
Your line “what would make you pull out your credit card right now?” is the right filter. I’d apply it backward to the 3 who actually did.
That small paid cohort is probably the product direction hiding inside the vanity metric.
500 signups with 0 paying is the classic validation trap - people say yes to free, not to value. What was the biggest insight on why they didn't convert?
The channel angle matters more than the positioning angle here. Twitter, HN, and dev slacks are tinkerer channels, they bring you curiosity, not budget. The actual buyer for b2b lead gen isn't browsing Show HN looking for tools. So 500 signups from those channels aren't a positioning failure; they're a wrong audience at scale failure. Even perfect positioning would have converted near zero in that crowd.
I'm in the middle of validating my own SaaS right now (different category, Android tooling) and reading this as exactly the trap i'm trying to avoid. Useful to see the cost of the alternative articulated this clearly.
Honest breakdown. 500 signups is a great signal, but the '0 paying' part is where the real stress starts — especially with AI.
We’re seeing this a lot: developers build something cool, get a surge of users, and suddenly they're drowning in API management and account limits while trying to figure out monetization. It’s why we started focusing on the 'plumbing' (account rotation, vaulting) with AiKey rather than just the feature set.
How are you handling the API usage for those 500 users while you iterate on the pricing? That burn can get scary fast.
How did you field users? What tactics did you use for publicity?
This resonates a lot. I'm in a similar spot — just built a competitor tracking tool for SaaS founders and trying to find my first paying customers. Did you try any outreach or just waited for organic signups?
One thing I keep noticing with AI SaaS tools is that signups are often driven by curiosity, not buying intent.
People think:
“this looks interesting, I’ll try it later”
—not:
“I need this badly enough to pay today.”
Especially in AI lead gen, where users have already seen a lot of similar tools.
Sometimes the biggest conversion issue isn’t traffic or even features, but whether the first session creates enough clarity and trust for someone to believe:
“this could realistically save me time or make me money.”
A lot of products lose people in that gap between signup and first meaningful outcome.
Curious what your onboarding looked like during those first 30 days.
You answered your own question, If you market to everyone you market to no one.
I'd suggest to refine your understanding on your specific market and ICP, and narrow your positioning towards it.
Other than that I can only reinforce what you've said yourself; Get feedback & iterate, and understand a prospect and a customer aren't the same thing...
There's some irony in the fact you're selling a lead generation tool, and it itself gets many leads but at a low conversion rate
Good luck man!
500 signups with almost no conversion is usually the moment founders realize attention and buying intent are completely different markets.
A lot of B2B AI tools can generate curiosity now.
Very few generate enough trust, specificity, or urgency to justify a payment decision.
“Lead gen for B2B companies” is broad enough that users can project anything onto it during signup — but when it’s time to pay, the positioning suddenly has to become concrete.
That’s also where naming starts mattering more than most founders expect.
Generic AI-growth style brands tend to convert curiosity better than commitment.
A tighter brand like Beryxa.com or Xevoa.com would fit this category much better if the product direction keeps moving toward serious B2B infrastructure rather than experimentation.
The 0.6% conversion rate isn't a product problem yet—it's a data segmentation problem. Those 500 signups likely fall into 3-4 distinct personas with completely different reasons for not paying (wrong ICP, didn't activate, pricing friction, wrong use case). Before rebuilding features, I'd pull your signups and tag each one: company size, acquisition channel, did they try the core workflow, did they return more than once. That segmentation usually reveals 1-2 cohorts where conversion should be high—and those are your 10 conversations worth having. I help SaaS founders set up this kind of customer data pipeline regularly. Also put together a free SQL diagnostic guide that covers segmentation queries you can run on your own DB → https://growthwithshehroz.gumroad.com/l/vgiex