Skip to content
German Orlov

How I work

I own the deployment from first workshop to measured adoption.

Four phases, each with the artifact it produces and the number it moves.

  1. 01 · Identify

    Frame the job to be done

    Requirements workshops with the people doing the work, not only the sponsor. Every candidate is framed as a job to be done.

    Artifact
    A ranked use-case shortlist, scored on value, data readiness, and regulatory friction.
    Number
    Baseline for the one metric the first use case has to move.
  2. 02 · Commit

    Design it, then fund it

    Solution architecture against the systems that already exist, plus a readiness assessment. The business case runs beside it. No number, no green light.

    Artifact
    Target architecture, integration map, and a business case with a named owner per workstream.
    Number
    Readiness scored across data, integration, legal, and the operations team that inherits it.
  3. 03 · Deploy

    Ship a thin slice with engineers

    One market, one journey, real traffic. I work inside the build with the engineers, not from a steering committee.

    Artifact
    A system live in production, with the runbook and guardrails that keep it there.
    Number
    The baseline metric re-measured, beside latency, containment, and error rate.
  4. 04 · Consume

    Make it stick, then scale it

    Adoption is a workstream, not an afterthought. Enablement for the teams who run it, then rollout market by market.

    Artifact
    Enablement material, an adoption dashboard, and a market-by-market rollout plan.
    Number
    Adoption rate, hours returned to the business, and the outcome the case promised.

The same model whether it's an agentic AI use case or a platform migration.

Building an AI platform, or getting agentic AI adopted at scale?

I work with teams turning AI and data strategy into systems that ship — and stick.

Start a conversation