DAG AIStart a conversation

And planned with your people at every step.

Not that one.

Independent advice.

We don't resell software. If the right answer is not to build, we'll say so.

What that means in practice
  • We tell you when not to build. Sometimes the right answer is a simpler fix, a tool you already have, or nothing at all. We'd rather lose the project than sell you the wrong one.
  • Realistic expectations. Set after we've seen your business.
  • You own what we build. The code, prompts, configuration, tests, and documentation we build for you are yours. Any team can pick them up.
  • Your data and systems stay yours.
  • Support with a clear scope and schedule.

Built for where AI is going.

AI keeps getting more capable. We build so you can hand it more as it earns your trust.

What that means in practice
  • Hand over as trust is earned. AI starts with one piece of work and proves itself there. Each success earns it more, and you set the pace.
  • You set the limits. You decide what each system can see, do, and approve.
  • Built to extend. The same foundations, tests, and patterns carry over as AI takes on more of the work.
The bar

Proven before you rely on it.

Nothing goes live until it meets an agreed bar on your real tasks.

What that means in practice
  • Results come first. Before any work starts, we agree what success looks like for your business and how we'll measure it.
  • Tested on your work. Demos are easy. We measure quality, cost, and errors on your real tasks, including the hard cases.
  • AI only where your team can trust it.
  • One measure for every option. A system that's cheap per call but needs constant checking isn't cheap. So we count everything it takes to get a result you accept.
Model and tool costs + people’s review time + reworkResults that met the barCost per accepted result. We compare every option this way, including the way you do it today.

Planned with your people.

Done right, AI makes work more human. As AI takes on more, people's work shifts and becomes more important than ever. We build for that from the beginning.

What that means in practice
  • Start with the people. Before we design anything, we sit with the people who do the work. We learn what wears them down and what they're proud of.
  • Bring everyone along. As AI takes on more, roles change. We plan that with your team in the open, and help each person grow into what comes next.
  • A plan for people whenever roles change.

How an engagement works.

  1. 01

    Talk

    A conversation about your business and where you want to go.

  2. 02

    Scope

    A written proposal with deliverables, a definition of done, and the cost.

  3. 03

    Deliver

    The work, done with your team and tested on your tasks.

  4. 04

    Keep improving

    Optional: ongoing checks and improvements on an agreed schedule.

Writing

Unsupervised Thinking

Trying to make sense of things that don’t.

Read on Substack
Hello.

Start a conversation

Tell us where you want to go.

If AI isn’t the right answer, we’ll say so.

Where would you like to start?

We’ll use your email to follow up about your interest. Privacy notice (opens in a new tab).