AI TRANSFORMATION SERVICES

We bring no menu: we learn what you have and build on it.

Made for one organisation: yours. We take a stream of work from the first idea to the last day it runs, with large language models, small language models, or a hybrid of both, and with no language model at all where plain algorithms do it better. Your KPIs are our service levels.

Talk to us

ONE METHOD, EIGHT STAGES

Decide. Prove. Build. Run.

Open each stage for what we do and what you hold.

  1. DECIDE

  2. PROVE

  3. BUILD

  4. RUN

  • DISCOVER AND ASSESS Consultancy first. A small team learns your business, the work worth changing, and what you already hold in people, data and systems, then ranks the opportunities with you. You leave with a plan and a clear go or no-go.You hold: a readiness report, a ranked portfolio with the value in each, and a go or no-go.
  • ARCHITECT We choose the technique the job needs: plain rules, machine learning, a small model you own, a large one, or a mix of them, and we record why each other option was set aside. Where it runs, how data moves and what it costs to own are settled here. This stage can be bought on its own.You hold: the architecture decision record, with every option rejected and why; a data and residency map; the cost of owning against renting; and whether one of our products fits.
  • PROVE Before anything is built, we show it can be done. A working version on a slice of your data, measured against a pass mark you set in advance. Pass, and you decide to build. Fail, and you have spent weeks, not a year.You hold: the proof report, pass or fail against criteria agreed in advance, an updated cost model, and the decision to build.
  • BUILD AND INTEGRATE This is our craft. We fit the model, build the software around it, connect it to your systems, and put a check before any agent acts. Independent testing before release.You hold: the working system, its documentation and the release evidence.
  • LAUNCH AND ADOPT Going live with your people, not around them. A phased release, the workflow redesigned with those who do the work, training, and a hand-over of the model, the tests and the documentation.You hold: trained people, adoption measures and the hand-over record.
  • RUN AND IMPROVE The Keep. One governance layer over every model, monitoring to the KPIs you set, and no change reaching production until your test has run again. Improvement is constant and systematic, for as long as the system serves you.You hold: a monthly governance and value report, the audit trail and every change record.
  • SCALE What worked is repeated across more workflows, sites and languages, reusing the method, the tests and the models. What did not earn its keep is retired.You hold: a scaled portfolio and an updated value ledger.
  • RETIRE AND HAND BACK When a system ends, nothing is lost. Its data, models, tests, decisions and lessons are preserved and handed to you, or carried into what replaces it. No copy stays with us.You hold: the exit pack, and the knowledge carried forward.

Every stage can be taken on its own, the architecture included.

THREE THREADS THROUGH EVERY STAGE

GOVERN AND SECURE

One inventory of every model, one audit trail, a gate before any agent acts, and the security controls from the first day. Aligned with recognised frameworks such as the NIST AI Risk Management Framework.

DATA

Readiness assessed at Discover, architected at Architect, governed from then on. Your records stay where they are, under your law.

VALUE

A ledger opened at Prove and reviewed each quarter, signed by your owner, so the value is measured, not assumed.

ARCHITECTURE IS NEUTRAL

We may conclude that no language model is needed.

At Architect we climb a ladder from plain rules, through machine learning and deep learning built in house, to search over your records, a small model you own, a large one, and agents only where a decision needs one. We stop at the first rung that passes your test. A product of ours is chosen only when it is the right answer, and we record why every other option was set aside.

THE KEEP

What the studio makes, the Keep runs.

Named for the stronghold where what you own is held: the models and products deployed for you, under one governance layer, measured against the service levels you set, for as long as you rely on them. The second half of AI transformation, the half most people forget to plan for.

YOUR KPIS ARE THE SERVICE LEVELS

Availability, latency, accuracy against your test, cost on your own hardware, time to resolve. We commit only to what we have shown is technically possible, and we show it before we sign.

EVERY MODEL, ALL THE TIME

Is it up, where you need it. Is it fast enough for the job, on the hardware it runs on. Is it still right, measured against your test as your records and rules change. What it costs to run, so there is no meter and no surprise.

ONE LAYER OVER ALL OF THEM

The inventory: every model, where it runs, what it does, who owns it. The gate: before any agent acts on your systems, a check you wrote and can read. The audit record: every question, every answer, every change, kept on your side. The review: a responsible-use review at a cadence you choose.

WHAT A MODEL UNDER USE CAN REVEAL

We watch for misuse of prompts and data, for attempts to pull a model off its job, and for changes in what reaches it. You receive what we see, in plain language, on a schedule we agree, and at once when it matters.

NOTHING MOVES WITHOUT YOUR TEST

No retraining, no rule change and no new model reaches production until your test has been run again and the result shown to you. Changes are recorded; records are yours.

SCALE WHAT WORKS

What earned its keep is repeated across more workflows, sites and languages. What did not is retired.

HANDED BACK WHOLE

When a system has done its work, or when you decide to stop, everything comes back: the model, the software, the test, the runbook and the record of every change. No copy stays with us. Confirmed in writing.

TWO WAYS TO HAVE IT

RUN WITH YOU

Your team operates. We stand behind: watching, advising, on call to the times we agreed.

RUN FOR YOU

We operate, inside your environment, to your KPIs. Nothing leaves.

THE PRODUCTS, RUN

What the studio made, the Keep can run: Key4MB across the phones your partner put it on, Key4KB on your server, Key4EDU on the school’s hardware, and the Key4 platform fitted to your job. The same inventory, the same gate, the same record, the same test run again before anything changes.

YOUR CLOCK

Coverage follows your hours, round the clock where the KPI needs it. When a contract needs people on your soil or more hands than we hold, we bring in partners we have worked with before, engaged for that contract and named in it.

The Keep earns its keep.

WHERE TO START

Enter where your need is.

Every engagement carries a tested exit.

  1. UNSURE WHERE AI PAYSDiscover and assess. A few weeks, a ranked plan, a go or no-go.
  2. A CLEAR PROBLEMArchitect. The technique, the data, where it runs, what it costs to own, and whether a product of ours fits.
  3. A PILOT TO RESCUEProve. Your test, your pass mark, a working version on your data, and a decision.
  4. SYSTEMS ALREADY RUNNINGRun. The Keep takes them on after an onboarding assessment.

THE FIRST ENGAGEMENT

Bring a problem. Bring an idea. Or pick one up.

One stream of work, chosen with you. The bar written down before we build, the test in your hands, and everything handed over whether or not you continue.

Talk to us

WHAT THE FIRST ENGAGEMENT LOOKS LIKE

  • ONE JOB One task, chosen with you, sized for a small model. Weeks, not quarters; the scope sets the time, and it is written down.
  • YOUR DATA The model is fitted on your records, on your hardware or in your room. Nothing leaves.
  • YOUR HARDWARE Usually nothing to buy. If the job needs a machine you do not have, we say so before we quote, not after.
  • AGREED BEFORE WORK STARTS Scope, bar and timetable, written down. No per-use bill on your own hardware, then or later.
  • A BAR, AGREED FIRST The accuracy and speed the job needs, written down before we build. You hold the test.
  • A MEASURED RESULT Against that bar, on your data, before you rely on it.
  • EVERYTHING HANDED OVER The model, the software, the test and the runbook are yours, whether or not you continue.

OWNERSHIP OF THE MODEL, THE TEST AND THE RECORD

No lock. Just the key.

Nothing we build ties you to us. You hold the model, the test and the record, and you can run them without us.

Let intelligence grow with trust, and all of us rise with it.