Your technology partner, not a phase of the project.

A strategy firm hands you a deck. A dev shop builds what the deck said. Neither runs it afterwards, and the one who does never built it. We do all four.

One of our people at a whiteboard covered in a hand-drawn process flow, working through it with the room
Four services

One relationship, at four points in its life.

Most people arrive saying one of these four sentences. Whichever one is yours is where the work starts. You do not have to buy the ones in front of it.

  1. 01

    Advisory

    You know AI should be doing something here. You do not know which thing.

    We trace how work moves, mark where people re-key data or hold a process in their head, and put the governance your board will ask for in place.

    • 4–6 weeks
    • Fixed fee

    Ends with A roadmap, and a working system already in production

  2. 02

    Custom build

    You know what to build. You want it built properly.

    Agents, the skills carrying your judgement, and connectors to what you already run. Plain files you can read and roll back.

    • Scoped to work
    • Fee or retainer

    Ends with Working software, licensed to you in perpetuity

  3. 03

    Implementation support

    It is built. It is not yet running where it needs to.

    Deployed into your infrastructure, credentials wired under your control, and your people taught to use it.

    • Scoped to work
    • Fixed fee

    Ends with It running in your environment, operated by your team

  4. 04

    Managed services

    It runs. Running it is not what your team is for.

    We host it, watch it, and keep it working as the systems underneath change.

    • Ongoing
    • Monthly

    Ends with Nothing. It continues, and you can take it in-house whenever

How an engagement runs

Discovery that ships something.

Most firms hand over findings and quote the build separately. We put a working system in front of your team before the roadmap is written, so you read the roadmap with the evidence beside it.

  1. Days 1–14

    Map how work moves

    We interview the people doing the work and trace it system by system, handoff by handoff. Every re-keyed field and every process living in one person’s head gets marked. Your AI policy and data guardrails are drafted here, in the opening weeks.

    A map, every gap named, and governance your board can review

  2. Days 15–28 Live from here

    Put something in production

    A working agent goes live in one function, answering against your real systems. No warehouse to build first, no migration. Your team is using it while the engagement is still running.

    Software in production, in one function, yours to keep

  3. Days 29–42

    Roadmap and readout

    What to build next, ranked by impact against effort. We present it to your leadership with the running system beside it, not a slide describing one.

    A sequenced roadmap, and a readout to leadership

What you get

Six things, and one of them is already running.

Everything below is ours to commit to and yours to keep. The fourth is the one nobody else hands over at the end of a discovery engagement.

  1. 01

    A map of how your business runs today

    Traced across every system in the path, with each handoff between them written down.

  2. 02

    Every gap, named

    Manual re-entry points, single-person dependencies, and the integrations that should exist and do not.

  3. 03

    AI policy and data governance

    Acceptable use, data handling, and a governance structure your board can review. Delivered in the opening weeks, not as a closing deliverable.

  4. 04

    A working agent, live In production, week four

    Answering questions against your systems, in production, in one function. Yours to keep whatever happens next.

  5. 05

    A prioritised roadmap

    What to build next, ranked by impact and effort, with a recommended sequence and a shortlist of quick wins.

  6. 06

    A readout to leadership

    Findings, the roadmap, and a demonstration of the running system to your leadership and board.

Who does the work

The engineers building your system also build the runtime under it.

We staff engagements with senior engineers and AI practitioners who work on the open-source runtime your agents run on. When your system needs something the platform does not do yet, they add it to the platform.

They are the people you meet on the first call, so the team that scopes your engagement is the team that delivers it.

A NimbleBrain engineer working through a wall of printed process diagrams

Proof

What it looks like when it is running.

14xMore concentrated outreach

Know Which Broker to Call Before You Send the First Email

A vendor tier rating told them who was in good standing. It did not tell them who places their kind of loan. So we scored for the second thing.

See the full story
A lender's desk in late afternoon light, a broker list annotated in pen beside a landline phone, with someone mid-call out of focus behind
The alternatives

What you are choosing between.

Four ways of splitting one job. Every seam in the row below turns into a conversation about whose fault something is.

How NimbleBrain compares with a strategy firm, a dev shop and a managed IT provider
Dimension NimbleBrain Strategy firm Dev shop Managed IT
Who scopes it The people who will build itA strategy teamYou, usuallyNot in scope
Who builds it Same teamNot in scopeThemNot in scope
Who operates it after Us, if you want thatNot in scopeRarelyThem, though they did not build it
Working software during the engagement Live in one function by day 28A roadmapAt the end, if the scope heldNot applicable
What you keep Everything, in perpetuity, on an Apache 2.0 runtimeThe documentPer the contractAccess while you pay
When the systems underneath change Covered under managed servicesNot in scopeA change orderA ticket
How it is billed Fixed fee, on milestonesTime and materialsTime and materials, or a fixed bidMonthly, per seat
Start with discovery 30 minutes. We will tell you if this is not the right shape for you.
A room mid-engagement, a process flow on the board and the team working against it

One team scopes the work, builds it, and keeps it running.

Before a first engagement

The questions that come up on the first call. Anything else, ask us directly.

How long does a first engagement take?

Four to six weeks is typical. Two to map how work moves, two to stand a working system up in one function, two for the roadmap and the readout.

Duration moves with the size of the business and the number of systems in the path. The three-part shape stays the same.

How does pricing work?

We agree a fixed fee before the work starts and invoice it against milestones rather than hours. You know the number in advance and it does not move unless the scope does.

We price it after a scoping conversation, because the answer depends on how many systems are in the path. Build and implementation scope to the work; managed services is monthly.

Do we have to start with discovery?

No. It is where most engagements start because most people are not sure what to build first, but the four services are separable. A team already running agents that wants them operated can start at managed services.

Do we have to move our data anywhere?

No. The runtime executes in your infrastructure and reaches your systems under credentials you hold. There is no warehouse to build first and no migration.

What goes live in week four?

One process, in one function, answering against your real systems rather than a prototype on sample data. We pick it in the first fortnight, out of what the mapping turns up, and it is the one where a person is doing the most by hand.

It is production software and it stays yours whether or not anything follows the engagement.

What happens when the engagement ends?

You keep a running system. The runtime is Apache 2.0, and everything we build for you is licensed to you in perpetuity for your business.

Some clients carry on into a build or managed services. Some run it themselves. Both were designed for.

Who does the work?

Senior engineers and AI practitioners who also work on the open-source runtime your system will run on. The people you meet on the first call are the ones who deliver the engagement.

We were burned by an agency before. Why is this different?

Three structural differences rather than a promise. The fee is fixed and billed against milestones, so there is no incentive to make the work take longer. Something is running inside the first engagement, so you are judging software rather than a plan. And the runtime is open source with everything licensed to you in perpetuity, so leaving costs you nothing you built.

Bring us one process that should not still be manual.