Labs

Journal

Field notes, not blog posts.

What we have seen work or break on real projects: search, AI answer engines, agents inside real companies, and software that holds up.

Articles

  1. GEO: how to get cited by AI answer engines

    Answer engines no longer rank ten links, they cite a handful of sources. What changes, what does not, and what we do about it in practice.

    • 5 min read
    • Search
    • GEO
    • Content
  2. AI agents in business: what works, what fails

    Field notes on AI agent projects: a narrow scope, clean data, human takeover, logs from day one — and the failures that keep coming back.

    • 5 min read
    • AI agents
    • Automation
    • Production

Playbooks

Short, applicable guides, written from our own projects and laid out with our own AI tooling. You receive them, you apply them without us — that is the point.

  • Playbook

    Getting an AI agent into production without getting burned

    The scope to write before any code, the data to clean, the human handover, and the logs to put in place on day one. With the checklist to run before going live.

  • Playbook

    Being cited by AI answer engines

    What GEO actually changes about search, what stays exactly the same, and the content trade-offs that decide whether you show up in a generated answer.

  • Playbook

    Framing a software project on one page

    The brief format we use: the problem, the users, what is explicitly out of scope, and what will count as success. It takes an hour and saves three weeks of misunderstanding.