FR Contact Menu
Contact

Method

The diagnostic decides. The solution follows.

One team, from the first interview to operations.

/01 · Diagnostic

How the diagnostic runs.

Nothing is estimated when it can be measured.

  1. 01

    Immersion

    The real work, desk by desk.

  2. 02

    Mapping

    Processes, data, systems, access rights.

  3. 03

    Measurement

    Volumes, lead times, errors, time spent. On real data.

  4. 04

    Prioritization

    Every case ranked: impact, difficulty, risk.

  5. 05

    Recommendation

    One solution per case. Sometimes no AI at all.

/02 · Solutions

When each solution fits.

And when AI is not the answer, the diagnostic says so.

  1. AI agents

    Cases handled end to end, inside existing systems.

    WhenMany varied cases, several systems to consult, decisions framed by rules.

    /0.1

  2. Automations

    Flows between systems. No more keying, no more rekeying.

    WhenRepetitive tasks with stable rules, between systems that do not talk to each other.

    /0.2

  3. Assistants

    The company’s knowledge, sourced, one question away.

    WhenKnowledge scattered across documents, frequent questions, answers that need sources.

    /0.3

  4. Custom software

    In place of generic software, a tool built around the company. And maintained for it.

    • ERP
    • CRM
    • Inventory management
    • Integrated finance
    • Customer portal

    WhenAn off-the-shelf tool imposes its processes, costs a lot per user or fits poorly with the rest.

    /0.4

  5. AI products

    AI at the core of applications and products.

    WhenAn internal application or a product gains a feature only AI makes possible.

    /0.5

  6. Process redesign

    Sometimes the best answer is a simpler process.

    WhenThe problem is the process itself: needless steps, unclear roles, unwritten rules.

    /0.6

/03 · Build

Autonomy opens step by step.

One more level, only once quality is measured.

  1. Level 1

    Suggestion

    The solution
    Prepares.
    The teams
    Decide and execute.

  2. Level 2

    Approval

    The solution
    Executes.
    The teams
    Approve.

  3. Level 3

    Supervised autonomy

    The solution
    Acts alone, within the set thresholds.
    The teams
    Handle the exceptions.

/04 · Engineers

An engineer embedded in the teams, from diagnostic to production.

The one who observes the work writes the code and puts it into production.

Example: the AI programme of an operations department. Five workstreams: a supplier matching agent and an order entry automation in production, a document search assistant in pilot, an inventory management tool being built and a customer reminders agent scoped by the diagnostic. Reviewed every week with the sponsor.

/05 · Run

Running over time.

  • /01

    Monitoring

    Quality, volumes, costs. Continuously.

  • /02

    Evaluations

    Every update replayed before production.

  • /03

    Models

    A new model only comes in if it does better.

  • /04

    Improvements

    New rules, new cases.

/06 · Ownership

Everything stays with the company.

  • /01

    The code, in the company’s repository

    Readable, tested, documented.

  • /02

    The evaluation sets

    Real cases, approved by the business.

  • /03

    The dashboards

    Volumes, quality, costs, business indicators.

  • /04

    The operating procedures

    When an agent hesitates, gets it wrong, or a system goes down.

Security and control

Contact