Article

Bitter lessons building AI products

Hex · hex.tech

A rare honest post-mortem from a product team on which AI bets worked in production and which had to be killed.

Open on hex.tech ↗

The question it answers

  1. Should my startup add AI features, and which ones actually matter to users?

Read with it

  1. A clear decision framework for the exact moment every founder faces: bolt on AI, ignore it, or rethink the product. founderprompts.com Article "Should we add AI?" Here is how to decide Vasil (Founder Prompts)
  2. The UX research heavyweights on why novelty chatbots harm products and what user value actually looks like. nngroup.com Article AI Features Must Solve Real User Problems Nielsen Norman Group
  3. VC-grade market data on which AI product categories are actually winning, so you bet where demand is proven. bvp.com Guide The State of AI 2025 Bessemer Venture Partners
  4. A blunt scorecard of which AI product ideas changed user behaviour and which quietly flopped. a16z.com Article State of Consumer AI 2025: Product Hits, Misses, and What's Next a16z
  5. Four consumer investors on why subtle product choices now matter more than raw model quality. a16z.com Podcast Where Does Consumer AI Stand at the End of 2025? a16z (Anish Acharya, Olivia Moore, Justine Moore, Bryan Kim)
  6. The canonical reframe: sell the finished work, not a 15% productivity tool, and your market gets 10-50x bigger. sarahtavel.com Article AI startups: Sell work, not software Sarah Tavel (Benchmark)
eChai Partner Brands The partnership →