Start a conversation

How we work

Every stage ends with a result you keep, and a decision you make.

An engagement is a series of stages. Each has its own scope, fee and written result. At the end of each one, you decide whether to go on, with us or with someone else.

  1. 1

    Assess

    How do the business and its systems work today?

    Assessment

  2. 2

    Plan

    What should change, in what order, and at what cost?

    Technology strategy

  3. 3

    Buy, build or both

    Which product, or what do we build?

    Software selection and implementation
    Custom software development

  4. 4

    Put into use

    Is it working, and are people using it?

    Part of both services in step 3

  • AI

    Alongside

    Where does AI pay, and how will we know?

    AI strategy

The engagement model

One stage at a time
Each stage has its own proposal: scope, deliverables, length, fee and the decision point at its end. You sign one stage at a time.
A first step with a fixed scope
An assessment, a review of a plan or proposal, an implementation review, an AI use-case assessment or a discovery for custom software. Each has a fixed scope and a fixed fee.
You can stop
After any stage you can stop, or continue with another provider, and keep everything the stage produced.

Decision points

Every service has decision points, shown here in the order they come up; not every engagement has all six. Your named decision owner decides. We recommend and set out the options.

Scope agreed
Before work starts: the questions, the scope and the fee.
Midway review
In longer stages: what has been found, and where to go deeper.
End of stage
Go on, change course, or stop.
Before a pilot
The success line and the stop line, agreed in writing.
Before go-live
Go or no-go, against criteria agreed at the start.
Handover
Who runs and supports what, and from which date.

Every decision goes into a decision log: the options, the evidence, who decided and when.

Who does the work

  • One lead runs your engagement from the first call to handover. The proposal and the contract name them, and if they have to be replaced, you agree the successor.
  • The proposal names everyone else who will do the work, specialists included, and those people do it.
  • Named people are replaced only with your agreement.

Working alongside your teams, vendors and integrators

Your teams
Your people score demonstrations, test scenarios and make process decisions. We write down what we learn, so it stays with you.
Vendors and integrators
We can lead the work, or act for you alongside them: reviewing estimates and change requests, and checking each stage against the plan before you approve it.
Your advisers
Legal judgement stays with your lawyers, and security testing with specialists. We work alongside both.

How we measure

  • Every number we report comes with its source and the way it was measured.
  • We measure processes and report the figures in aggregate. Where employee representatives, such as a works council, have a say in how work is measured, the method is agreed with them before we start, and we prepare what they need to decide.
  • Results are measured the same way as the baseline, so before and after can be compared.

How fees work

  • Fees are agreed before each stage starts. Assessments, strategies, selections, reviews, discoveries, AI use-case assessments and pilots have fixed fees. Implementation is priced by milestone. Building custom software is priced by release, or by time with a monthly cap. AI in production is priced by release or milestone.
  • What drives the fee: the number of companies or business units, systems, countries and integrations in scope, and the amount of data to move.
  • Currency, payment terms and the expenses policy are in each proposal.

Confidentiality and your data

  • We sign a non-disclosure agreement, yours or ours, before you share anything sensitive.
  • We ask for read-only access wherever it is enough, and use your accounts under your policies.
  • We take the minimum data the work needs, and anonymised samples where they will do.
  • At the end of the engagement, working files are returned or deleted, as the contract sets out.

How we use AI in our own work

We use AI tools in our own work, and we tell you how.

Where
Drafting and editing documents, summarising interview notes, searching large sets of documents, analysing data samples, and writing and reviewing code.
Who answers for it
A named person checks every output before it reaches you, and answers for it as if they had written it alone.
Your data
Client information goes only into tools whose terms forbid training on it, and only with your agreement. At the start of an engagement we list the tools we will use, and you can rule any of them out.
Figures
Every figure in our reports traces to its source data.

What we don’t do

  • Day-to-day IT operations, hardware supply and round-the-clock support. When a project depends on them, the plan says so, and we help you find a provider.
  • Security testing, formal audits and certification. We flag where they’re needed and help you scope them.
  • Legal advice. Contract terms stay with your lawyers.
  • Training our own AI models. We use established models and build what surrounds them.
  • Hosting or operating your production systems.
  • Acting as main systems integrator on programmes that need hundreds of people. On those, we work for you alongside the integrator.