Field notes · The Wayfinder Notes

How AI-native companies actually get built.

I'm building CURN and writing down what the work is teaching me: how teams work with agents, how useful workflows become memory and where the advantage actually settles. For founders and operators building with agents in the loop.

Start with a field note ↗ Browse the log →

The question

Good AI work should not disappear when the chat closes.

One person can suddenly do far more, while the rest of the team sees far less of how the work was done. Useful prompts, decisions and workflows disappear into private sessions, and the next person starts again.

Once the work is captured as shared memory, one person's useful method becomes something the whole team can build on. The Wayfinder Notes is where I write down what I'm learning from inside the build. You should leave each piece knowing what changed, who wins and what to look at next.

Topics

When agents enter the operating model, the company changes with them. These are the frameworks I work from.

  1. The AI Demonstration Gap The gap between what an organisation says about its AI capability and what its people can demonstrate in real work. Most AI investment dies in the distance between the two.
  2. The Willingness Gap A genuinely useful tool goes untouched because the person at the desk gains too little from using it. Where value runs aground.
  3. AI adoption Not tool rollout. How isolated AI wins become work a team can see, repeat and trust.
  4. Team memory How good AI work escapes the private chat and becomes something the team can reuse: the record of context, decisions and lessons that should change the next run.
  5. Strategy signals How to read weak signals before they harden into obvious market shifts.

Latest entries

The full log →

Building

CURN ↗

CEO

Where I test the thesis against real work: turning good AI work into playbooks the rest of a team can actually use, instead of leaving it trapped in private chats.

Behind the notes

James Kerr

CEO of CURN

I'm building CURN and writing The Wayfinder Notes from inside the work: AI adoption, agent workflows and reusable team memory.

Next entry

Useful work should compound.

I'm writing down what I learn as I build with agents in the loop. If you're working through the same questions, I'll send the next field note straight to your inbox. Get the next field note ↗

Connect

Longer pieces go to Substack. I continue the conversation on LinkedIn and use X for rougher ideas that are not ready to settle yet.