Field Notes
More notes on AI search, AI agents, automation, web, client portals, platforms and growth, from the people who build them. This is page 2 of 5.
Which pages are worth writing for AI search
You do not win AI search with more pages, you win it with the right ones. Which pages to write, and which quietly waste your time.
How to tell if your GEO is actually working
GEO has no neat ranking report. How to actually measure whether AI engines name you, and what real progress looks like.
Where AI agents still fail, and how we design around it
Agents break in predictable ways. Where they fail, and how grounding, guardrails and handoff catch it before your customer does.
How we stop an AI agent from making things up
What if it just makes something up? A fair fear with a real answer: how we design honesty into an agent.
Why automations break over time, and how to make them last
An automation that works today can quietly stop working next month. Why they break, and what makes one durable.
The manual work quietly costing you the most
The costliest manual work is the small task repeated all day. How to find yours, and why the fix is usually small.
The hidden cost of a cheap website
The cheapest quote wins a lot of projects and loses money after. What a cheap site really costs you, where the quote is silent.
What clients actually want to see in a portal
Most portals miss the few things clients came for. The three things they truly want, and why less is more.
Signs you have outgrown your spreadsheets
Spreadsheets are brilliant until they are not. The signs one has quietly become critical, and when a real system pays off.
The few metrics that tell you marketing is working
Most dashboards are full of numbers that decide nothing. The few that actually tie effort to money, and the rest to ignore.
Rather talk it through than read about it?
Tell us the problem in a couple of sentences. You will get an honest answer on where to start, or whether we are even the right team for it.