I’ve built my AI startup idea into a real product, but now I’m stuck. I need honest feedback.
Hey everyone,
I’m a solo technical founder, and I could really use some advice from people who have been through this stage.
Over the past few months, I’ve been building an AI startup idea called Ainexa.
I started with the idea of helping AI builders discover promising opportunities before spending months building the wrong thing.
So I built:
an AI opportunity radar
market research reports
opportunity scoring
analysis of trends from GitHub, Product Hunt, Reddit, and other signals
The product is live, and I’ve spent a lot of time building, deploying, and trying to share it.
But here is where I’m stuck:
I’m not sure if I’m solving a real painful problem.
The feedback I’ve received is that the idea is interesting, and people like the concept of "turning AI market noise into build decisions."
But I still don’t have enough evidence that founders would actually use it regularly or pay for it.
Right now I’m questioning:
Is this the right problem?
Should I keep pushing this direction?
Should I change the positioning?
Or should I move on to something else?
For indie hackers who have built SaaS products:
How did you know when to keep going vs. when to pivot?
How did you validate that your product solved a painful enough problem?
If you were in my position, what would you do next?
I’m not looking for encouragement — I’m looking for honest feedback, even if the answer is "this is not a good business."
Thanks to anyone willing to share their experience.
Hi @miniss02589, thanks for your thoughtful feedback. I really appreciate it.
I agree that positive feedback is very different from real user demand. I’m trying to validate the problem more carefully and focus on specific workflows instead of just collecting general opinions.
By the way, I’m currently having an issue with my Indie Hackers account. I can see notifications and replies, but I can’t open some links or access the discussion properly. It seems my account permissions may have been limited, so I’ve contacted Indie Hackers support to check.
Thanks again for your advice!
I’d probably avoid making the keep-or-pivot decision based on general positive feedback. “Interesting idea” is very different from “I need this.” I’d narrow the target user down as much as possible and have 10–15 conversations around one specific workflow. If the same pain keeps coming up without you leading the conversation, that feels like a much stronger signal than feature feedback.
Circling back since it has been a few days, in case my last comment got buried under the thread. Still keen to trade a look whenever works for you. Just drop a day this week plus your timezone here, or an email if that is easier, and I will set up the FounderFlow walkthrough around it.
Hey, thanks for following up!
I’d be happy to take a look at FounderFlow and exchange feedback.
Saturday 10 PM (UTC+8) works well for me. Let me know if that works for you, and feel free to send the meeting details.
Looking forward to seeing FounderFlow!
Got it, Saturday 10 PM UTC+8. Let me lock that in on our end and I will send you the meeting link and details shortly. Glad this is finally happening.
Sounds great, looking forward to it!
Thanks for setting this up. I’m excited to exchange ideas and discuss how AI signal discovery and decision-making systems can complement each other.
See you Saturday. 🚀
There is no direct messaging here on Indie Hackers, so easiest way to get you the actual meeting link and details is by email. Could you drop your email address here, or send one to hello@founderflowhq.ai? I will get Saturday 10 PM UTC+8 locked in on our end and reply with the link.
Thanks for reaching out! I’ll send an email to hello@founderflowhq.ai. Saturday 10 PM UTC+8 works for me. Looking forward to the conversation!
Got it, thank you. Saturday 10 PM UTC+8 is noted on our end. Once your email comes through I will confirm it against the team's calendar and send you a real link, not a placeholder. Talk soon.
Hey, just checking in. Are you still looking for feedback on your product? Happy to take a look whenever works for you. 🙂
Thanks for the follow-up!
Saturday 10 PM UTC+8 works for me.
I’ll send over my email so you can confirm the calendar invite.
Looking forward to the conversation and learning more about your perspective.
This comment was deleted a month ago
This comment was deleted a month ago
The signal to watch for is not whether people like the concept, it is whether anyone changed an actual building decision because of a report your tool produced. Interesting and worth paying for regularly are very different bars. Try asking the people who gave you positive feedback if they would be upset if the product disappeared tomorrow. If the honest answer is no, the problem is not painful enough yet, even if the idea itself is sound. I would not throw out the whole direction yet, I would narrow it first. Pick the one signal source, GitHub, Product Hunt, or Reddit, that has produced the most useful opportunity for someone so far, and go deep on serving that single case well before trying to cover all four at once.
Thanks for the thoughtful feedback. I think you pointed out the key distinction I was missing: people being interested in the concept is not the same as the product creating enough value to change someone's decisions or become part of their workflow.
The “would you be upset if the product disappeared tomorrow?” test is a really useful way to think about real demand. I agree that I need to focus less on collecting more positive signals and more on finding users who would actually rely on the insights to make better building decisions.
I’ll narrow the scope and focus on one signal source first, especially looking for places where users have strong pain or are actively searching for solutions, rather than trying to cover everything at once.
Really appreciate the perspective — this is exactly the kind of feedback I was hoping to get.
Thanks for the thoughtful feedback. I think you pointed out the key difference I was struggling with: people finding the idea interesting is not the same as the product creating enough value for them to change decisions or pay regularly.
The “would you be disappointed if it disappeared tomorrow?” test is a very useful way to think about this. I agree that I need to focus less on collecting more signals and more on proving that the signal actually helps someone make a better building decision.
I’ll narrow the scope and test one specific signal source first, rather than trying to cover everything at once. I appreciate you sharing this perspective — this is exactly the kind of honest feedback I was looking for.
That is a good instinct. One thing that helped me when narrowing scope is picking the signal source based on who complains loudest, not which one is easiest to pull data from. A source that is technically convenient but emotionally low stakes for the user will not move anyone to pay, even if the data quality is great. Good luck with the test, curious to hear what you find.
Thanks, this is a very valuable perspective.
I think you’re right that I should pay more attention to where users feel the strongest pain, rather than simply choosing sources that are easier to collect data from. A technically accessible signal does not necessarily mean it creates enough value for someone to act on it.
This also reminds me that the real validation is not whether the tool can find interesting opportunities, but whether it helps someone make a better decision or avoid wasting time.
I’ll narrow the focus and look for the signal sources where people are already actively struggling or looking for solutions. Appreciate you sharing this — I’ll keep you posted on what I find.
Thanks, this is a very useful perspective.
I was initially thinking more about which signals I could collect, but your point changes the question: the source should be chosen based on the user's pain level, not just data availability.
I’ll focus more on finding where founders have the most uncertainty and where a wrong decision actually costs them time or money. Appreciate the insight — I’ll share what I discover from the test.
Sounds like a solid next step. Would be curious to hear whether the pain level signal ends up correlating with willingness to pay, that gap is where a lot of validation tests fall apart even when the interest is real.
Hey Stacy, thanks for your thoughtful comment on my post.
Your point about the gap between pain signals and willingness to pay really resonated with me.
I noticed you’re building FounderFlow around helping founders avoid losing opportunities buried in communication. That problem feels closely related to what I’m exploring — helping founders identify better opportunities before spending months building.
I’d love to get your perspective as a founder. Would you be open to taking a quick look at what I’m building and sharing your honest feedback?
Hey FounderFlow_57, thanks again for the thoughtful feedback.
Your point about the gap between pain signals and willingness to pay is exactly what I’m trying to validate.
I’m building an AI market opportunity radar for founders, and I’d love to get your honest perspective since you seem to have experience thinking about validation.
Would you be open to a quick chat? If easier, feel free to share your preferred contact (X/LinkedIn/email).
Happy to take a look at Ainexa, that opportunity-scoring problem is a real one. Fair trade since you already asked about FounderFlow: it's built for founders tracking their business by memory and spreadsheets instead of one system, and I'm still onboarding the first thirty founders personally. Want to trade a look? You show me Ainexa, I show you FounderFlow, we both get real feedback instead of guesses.
That sounds great — happy to trade a look.
I’ll show you Ainexa and would especially love your feedback on whether the opportunity scoring actually helps you make a different decision, not just whether the interface or idea looks interesting.
I’d also be happy to take a look at FounderFlow and share feedback from a founder perspective. Let’s exchange notes.
Deal. Send over the Ainexa link whenever it's ready and I will go through it properly, not just skim it.
For FounderFlow I would rather walk you through it live than send screenshots, since the useful part is seeing how it flags things, not just the interface. What does your week look like for a short call? I can work around your schedule.
Sounds great, thanks!
I’m finishing up the Ainexa prototype now. I’ll send you the link once it’s ready.
I’d also be happy to go through FounderFlow with you. A live walkthrough sounds actually better — I’m interested in understanding how founders interact with the decision process and what signals the product focuses on.
Looking forward to trying it out and exchanging feedback.
Good, a live walkthrough it is. Since we are both heads down building, easiest way to actually land on a time is to trade emails instead of going back and forth in comments. What is a good email for you, or if you would rather keep it here, just give me a day this week plus your timezone and I will work around it.
Finding founders who already paid for a wrong decision is the right filter. Pain that shows up as real dollars lost is very different from pain someone will nod along to in a survey. Curious what channel you are using to find people who have that story fresh, because that is usually the hardest part to source.
Thanks, I agree — the hardest part is finding founders who have actually paid the cost of a wrong decision.
I'm currently looking in Indie Hackers and other founder communities. My goal isn't to collect opinions or validation surveys, but to find founders who recently made a bad bet and see whether AI analysis can help them reconsider their direction before spending more time and money.
If you know any founders with that experience, I’d really appreciate an introduction. Happy to chat here or via DM if that’s easier.
Exactly. That’s probably the most important thing I’m trying to validate.
A lot of people say a problem is painful, but the real question is whether that pain is strong enough to change a decision or create a budget.
I’m going to focus on finding founders who have actually paid the cost of making the wrong decision — wasted months, money, or resources — and see if better signals would have changed their actions.