I shipped an AI chatbot for websites. 0 customers. Here's what I've built and what I'm doing wrong.
Hey IH. Build-in-public update that's more "build-in-humility."
What I built: Factfu — an AI chatbot that reads your website/store content and answers customer questions. RAG-based (retrieval augmented generation), so answers are grounded in real data, not hallucinated.
Features shipped:
Website crawling + document upload
Shopify product sync (auto-imports entire catalog)
Live chat with human handoff
Lead capture forms
Slack + Zapier integrations
Custom branding (colors, fonts, position)
Order tracking (Shopify stores)
Pricing: Free (200 msgs/mo) → $19/mo (2K msgs) → $49/mo (5K msgs)
Revenue: $0. Customers: 0. Time invested: months.
I know the product works — I've tested it extensively. The problem is 100% distribution. Nobody knows this exists.
What I'm doing now:
13 SEO blog posts targeting "add AI chatbot to [platform]" keywords
SaaS directory submissions
This post
What I'm asking you:
If you have a website with customer questions, want to try it for free? factfu.com
What distribution channels have worked for your dev tools/SaaS?
Honest take — is this market too crowded or is there room?
Not looking for validation. Looking for signal.
Checked your website and product. The website looks very generic and obviously AI-generated. The product still needs to be improved - it can't answer simple questions like "1+5", also it doesn't add more information when waiting for reply. Good luck!
Thanks for trying it out! On the "1+5" question — FactFu is specifically designed to only answer questions grounded in the business's actual content. So if a store doesn't have math in their FAQ, it won't make up an answer. That's the core value prop: zero hallucination, only answers backed by real data.
Fair point on the website — I'll tighten up the design. Appreciate the honest feedback.
This is a great breakdown. Feels like one of those cases where the product works, but distribution + differentiation are the real bottlenecks. If you had to pick one type of user (Shopify store owners, SaaS founders, agencies), who’s seen the most value so far?
Spot on — distribution and differentiation are exactly where I'm stuck. Honestly, I haven't had enough users yet to know which segment gets the most value. My hypothesis is small business owners who can't risk a chatbot hallucinating wrong info about their products/policies. But that's still an assumption, not data. Open to suggestions if you've seen what works for getting those first users for a tool like this.
That makes sense and I like that you’re being honest about it still being a hypothesis.
If I were in your shoes, I’d pick one very specific segment (like Shopify stores with frequent order/support questions) and go deep there instead of broad “small businesses.” It makes messaging and distribution way clearer.
For first users, I’ve seen a few things work well:
– Reaching out directly to store owners already complaining about support load
– Offering to set it up for them for free in exchange for feedback
– Positioning it around a clear outcome (e.g. “reduce repetitive support questions”)
I’ve been working with a team at Creatives Takeover where we help founders with early validation + getting those first users, especially at this stage. Happy to share more if useful.