> It remembers
everything that gets decided, who said it and when.
Have you talked with someone inside a company? I agree this is a problem, but I'm not sure it is a problem people want to solve (and are willing to pay for it).
Impressive landing page, but requires too much time and attention to parse. The typical business decision maker isn't going to waste their time with it.
You need to accurately and succinctly express the benefit to earn the attention for them to explore further.
Outside of tech circles, AI is viewed with suspicion due to over-inflated claims. No business person buys AI to be seen as being cool. They will buy a product that delivers tangible benefits -- increased profits, reduced costs.
They buy AI to get investors to increase the company's stock price for being up with AI. That is a tangible benefit, but probably not the kind you were thinking of.
Hey HN. I'm 16, in high school in France. I was in TKS (The Knowledge Society) for a couple of months, did a few online hackathons, and since May I've been building one thing on the side, around 40 hours a week on top of school. My brother runs the company, I build.
It's called Dermioz. Basically an AI for companies that remembers who decided what and when. The thing that bugged me: AI at work just answers from whatever text looks closest to your question, so if an old price is repeated in five meeting notes and the person who changed it said it once, the old price wins. In mine, what a person decided beats a document, no matter how many times the doc repeats the old value. In our test it kept the right answer 60/60 where a plain vector search only did 6/60.
It's not perfect. Synthetic data so far, no paying customers, first test in a real company in January, and the model is an open one, not ours. I wrote it with AI tools, the choices and the tests are mine.
Curious what you'd need to see before trusting something like this at your job. Roast it if you want, I'd rather hear it now.
Have you talked with someone inside a company? I agree this is a problem, but I'm not sure it is a problem people want to solve (and are willing to pay for it).
reply