I built competitive inteligence tool at my job, watched it die, now productizing it – looking for ~10 testers
I work in product. Recently, at a company I worked at, I ended up building a competitive comparison tool myself because nothing else existed. It was useful for about a month – sales actually used it a lot – then it went stale because keeping it updated wasn't really anyone's job.
What made me actually build this instead of just complaining about it: I've seen the same pattern at more than one company I've worked at. Different product, different market, same result – someone builds a doc, it's great for a month, then it quietly rots because keeping it current isn't anyone's actual job.
So I'm rebuilding it as the version that doesn't die. You map your competitors and your own product's features together, AI helps fill it in, and from there it keeps itself alive in the background – flagging when something on a competitor's side actually changes. It's turned into something broader than just a competitor tool too – a place where your own product is written down clearly, which matters just as much when someone asks "wait, do we actually have X?"
This is a first version I've built myself, mostly to validate the idea and get real feedback – so I'm quite early stage here, not polished.
Good fit if:
-You already track competitors somewhere (Notion, spreadsheet, memory) and it keeps going stale
-Or people keep asking you about your product's features/limits and there's no clean source of truth
-You've got a couple of real competitors or comparable products you could load in
Probably not a great fit if you're pre-launch or nobody's asking these questions yet though feedback on the tool itself is always welcome regardless.
Free, no card, no strings. Looking for ~10 people to load real competitors in and break it. I just want honest feedback on what's confusing or missing.
"Rots because keeping it current isn't anyone's job" is the whole product, and it's your validation, you saw the pattern at multiple companies. Lead with the death, not the tool.
The make-or-break: everything except auto-update is table stakes (Notion does it). So the magic isn't detecting change, it's detecting change that matters. Flag every trivial price test and word tweak and you're noise, people tune out, and that's how the manual doc died, just automated.
What's your signal-vs-noise approach on change detection? That's the whole ballgame.
This is the sharpest comment I've gotten on this anywhere so far, thanks :)
You're right, and it's the part I don't have fully solved yet. right now I'm doing it in two layers – cheap check first (did the page even change at all), then only if yes, a closer pass to see if the change is actually meaningful (price, plan, feature) versus a wording tweak or an a/b test. the second layer is the part I'm least confident in. it's the difference between the tool being trustworthy and it being the same dead doc, just automated, like you said.
what would you consider "meaningful" if you were setting the bar? price and plan changes feel obvious, feature/changelog stuff is murkier.
A bar that cuts through it: a change is meaningful if it would change what a sales rep says in a live deal. That's the whole test. Price, plans, a feature that closes a gap you used to win on, a limit removed, a new integration, all change the answer to "how do you compare to X." Wording tweaks, design refreshes, A/B copy don't. If it wouldn't alter a battlecard, it's noise.
That also solves your murky changelog case: an entry matters only if it moves a competitive line (they shipped the thing your customers kept asking for, or killed something you cited as your edge). The classifier question isn't "did the feature set change," it's "did the competitive position change."
Practical filter: anything touching a row in the user's own comparison table (price, features, limits, integrations) is meaningful by definition; anything off that table is probably noise. They already told you what matters when they built the table.
And bias to precision over recall early, one false "competitor changed!" that's really a button color and users stop trusting every alert, back to the dead doc.
This competitive-positioning-signal stuff is exactly what I spend my time on, part of the team building Hivemind, an AI strategy copilot. Might be a useful reference for how to frame "what's a meaningful competitive move": https://hivemind.myosin.xyz
Good filter – "does it change what the rep says" maps cleanly onto the comparison table itself. Appreciate the framing.
Congrats on productizing it! For competitive intelligence, tools like Ahrefs, Semrush, and Similarweb are useful, but they can get expensive and overwhelming. SerpSpur takes a simpler approach with an agency-style website audit covering technical, on-page, off-page, local SEO, and actionable issues—useful for founders who want clear next steps rather than just data.
The internal tool dying is useful evidence if you can identify why: no owner, no budget, weak integration, or insufficient urgency. Before productizing it, interview the people who were supposed to act on the intelligence and test one decision the tool can improve repeatedly.