How we work

What working with us looks like

We work in the same three steps with software companies, with property managers, lodges and short-term rental owners, and with leaders whose AI runs on their data, and you see the return in your own numbers.

Working together

The three steps

Before we start, we agree what will be live within ninety days. If it isn’t, we keep building at no charge, as long as we’ve agreed the scope and have access to your systems in time. Live means in use by your people, in your own systems.

One fixed price, agreed before we start, covers finding what pays and building it in, through go-live. Keeping it running is optional, month to month.

  • Find what pays We listen to you, your leaders and your team, and pick with you where AI pays off first.You come away with the improvements we’ll start on and the measures we’ll judge them by, each with a baseline from your own systems, or a first reading where no history exists.
  • Build it in We make the improvements with your people, in the systems you already run.You come away with the improvements live, and you see how far each measure has moved so far.
  • Keep it runningIf you choose After that, you decide whether we keep going alongside your team as your systems and the AI providers’ models change.Either way, it runs in your accounts, documented, so your people or anyone you choose can keep it running.

A normal week

Our founder, Jeremy Brooks, meets with you and your leaders regularly on the problems and where the business needs to go, and you stay out of the detail.

Between those meetings, we work with your team, turning what we learn into action and bringing it back to you.

Your team stays in charge

We work with every layer between you and the people doing the work rather than going around anyone, and whoever leads that work stays in charge. The team is our founder, senior builders like Patrick Hrabos who join when the work needs them, and the AI agents we work alongside. What your people tell us goes to their leader first, and nothing about any one person goes up the chain.

  • We start from what already works, why it was built that way, and the chores your people most want to give to AI agents.
  • Nothing changes how your people work until their leader says yes.
  • Everything we create, from code to prompts and agent settings, lives in your systems and belongs to you, and our agreement says so.
Who stays in charge, and what stays yours

For software teams

The Four Handoffs

Engineers report working faster with AI, and too often that never reaches the customer. DORA, Google’s research program on software delivery, says in its 2026 report on AI returns that without the right organizational foundation, “AI creates localized pockets of productivity that are often lost in downstream chaos.” Those gains leak out at the handoffs, the points where work passes from one person or team to the next, and AI puts four of them under new strain.

Here is what we work toward at each one, each with a measure your CTO agrees on:

  • Intent, strategy to the people building. Engineering time goes to your top goals, and you can see how much each one gets.
  • Ownership, decision to owner. Less work waits on decisions, because everyone knows who can say yes.
  • Review, AI output to trusted code. Fewer mistakes reach your customers, and finished work reaches them sooner.
  • Learning, one engineer to the whole team. What one engineer learns about working with AI reaches everyone, so the same mistakes repeat less.

The engineering side shows how each one works and how we measure it.

The Four Handoffs in one minute

What the business needs

  1. 1 · IntentStrategy to the people buildingThe plan changes on the way down.
  2. 2 · OwnershipDecision to ownerWork waits while nobody is sure who can say yes.
  3. 3 · ReviewAI output to trusted codeCode gets written faster than anyone can review and trust it.
  4. 4 · LearningOne engineer to the whole teamThe gains stay with one person.

Results your board sees

For software teams

How we measure it

Your board will ask three things, and the measures build on DORA’s delivery research and on DX’s frameworks for developer productivity.

  • Are we building what the business needs? Progress on your top two or three goals against their baselines, how much of what engineering commits to gets delivered, and how much engineering time goes to each goal and to new capabilities.
  • Is delivery faster without getting riskier? Time from idea to customer, how long work waits on review, how often a change fails in production, and how many deployments are unplanned fixes.
  • Is the AI spend paying back? AI spend and use by team, set against those results, with any self-reported time savings checked against what actually shipped.

Everything is reported by team and never by person, with no leaderboards. Numeric targets go on your business goals and never on the delivery measures, because a measure that becomes a target stops telling you the truth, as Goodhart’s law warns. Early numbers can dip while a team changes how it works, which DORA’s 2026 report on AI returns expects, so the readout you take to your board reports how far each measure has moved, including any dip. Every number in it is marked as measured, estimated with a range, or pending. Our founder wrote more about what to bring when your board asks if AI is paying off.

For software teams

What you take to your board

  • Answers to the three questions above, in the same order, each set against where you started, in your own numbers.
  • What changed in each handoff we worked through, and what your team now owns.
  • Your CTO presents its engineering side.