CharlieKey takeaways from your favorite podcasts
← 6 October 2026The Startup Ideas Podcast

13 Businesses for the Age of AI

In an agent-driven world, the host argues only 13 kinds of businesses are worth owning, from AI-native service firms and distribution to harnesses, physical assets, compute and security.

Watch on YouTube ↗

Charlie's summary

  1. AI-native service firms: agents do about 90% of delivery, humans review exceptions, and the client pays per completed job

    The host says an AI-native firm doesn't need agents to do all the delivery; 90% is enough. His example is a bookkeeping firm for one type of business: agents prepare the books, a human reviews exceptions, and the client pays monthly or per completed job. He likes them because they are high margin and need few clients. He is actively ideating on one, such as logos or ads as a service.

    ▶ 1:01
  2. Offline businesses win because people still pay for shared experiences, and you should start with the food-truck version

    The host points to Othership, a communal sauna club with membership, as proof that people pay for wellness and community; agents can handle admin, but “the room, the experience is what people are buying.” Offline is costly, so he asks what the food-truck version of your business is. Othership's first version was literally in the founder's backyard, then friends and community, then a space.

    ▶ 2:39
  3. Distribution is the scarce asset, so use the ACP funnel: audience, community, product

    Building products is easy now; reaching people is harder, and the host expects it to keep getting pricier. Facebook ads cost about 5 cents a click at launch and now run a dollar to several dollars. His answer is owning organic audiences. He calls the sequence the ACP funnel: build an audience, then a community, then build the product they ask for. Advertisers can even pay you while you build the audience.

    ▶ 5:21
  4. Proprietary data sets are a gold rush, and we're in the first inning

    The host says data is the new oil. Companies like Merch One and Micro One sell data to the big labs at billion-dollar valuations. A solo founder can aim smaller: a business making a few million a year from useful information people can't get from their personal agents or ChatGPT. Owning data the models lack is a strong position.

    ▶ 7:30

Give your email to unlock the rest of the key points — for every episode.

Do more with Charlie

Free — sign in with your email, no password.