Draft for Tamir's review. Not published.
What to check
Start with the claims in the pitch: what the product does, how it scales, and what is proprietary. Then ask for the evidence behind each one, such as a live system, usage data and the code.
For AI products, separate what the company built from what it rents. If the core runs on a third-party model, check the cost per use and the terms. Model APIs are priced per million input and output tokens, and prices change often (source: Anthropic pricing page).
Risks a demo won't show
Look at who knows the system, how much depends on one or two people, and how releases reach customers. Check who owns the code, the data and the AI output.
Check the regulatory calendar for the markets they sell to. In the EU, AI Act high-risk obligations apply from 2 December 2027 for the listed areas (source: European Commission).
What you get
A written opinion on whether the technology matches the claims, where it doesn't, and what it would take to close the gap. The investment decision stays with you.
Asking me is free. You get my initial view and the question I think matters most. A full review runs under a short agreement, and the first one includes a no-value, no-fee clause.
What to check before you decide
- List every technical claim in the pitch and ask for the evidence behind each one.
- See the product run on real customer usage, beyond a prepared demo.
- Ask which parts depend on third-party models or services, and what each costs per use.
- Check who owns the code, the data and the AI output under the contracts in place.
- Ask how many people can change and release the core system without help.
- Check which AI and privacy rules apply to their customers, and from when.