01
Focused diagnostic
Used to clarify a use case, a flow or a friction point before anything is built.
- business framing,
- technical and data constraints,
- risks, assumptions and priorities,
- execution-ready framing note.
Engagement models
Short, well-framed and directly useful engagements.
The right format is not the one that promises the most. It is the one that clarifies the subject quickly, produces a readable deliverable and leaves a clear next step if the case deserves follow-through.
Formats
Four client-readable engagement shapes
Each format is designed to produce a readable result, not just exploration or idea backlog.
02
Useful prototype
Used to make an assistant, workflow, copilot or data layer visible quickly.
- testable prototype,
- first concrete outputs,
- initial measurement frame,
- recommended next steps.
03
Operational pilot
Used to run a real workflow on a limited perimeter with human control and explicit ownership.
- pilot system or workflow,
- guardrails, observability and stopping points,
- business feedback loop,
- pilot memo and next-step logic.
04
Selective hardening
Used to strengthen what genuinely deserves to become more reliable or more broadly usable.
- evaluation and quality,
- readiness note,
- metric layer or control logic,
- clean handoff into the next stage.
Fit
When this model is the right one
Good engagements start from a readable situation. Judgment is part of the value proposition.
Good fit
What works well with this style of engagement
- a business need is already identifiable,
- the goal is to move quickly on a limited scope,
- the flow, documents or data already exist,
- the team needs a first system, framing note or clean pilot.
Less good fit
What usually calls for a different setup
- a large programme is still too vague,
- the expectation is a full platform from day one,
- the need is a large full-time implementation team,
- the work is pure academic research without a target use case.
Typical deliverables
What a client should reasonably expect to receive
The objective is to produce outputs that support decisions, not simply more slides.
Executive note
Framing, decision points, risks, trade-offs and recommended next step.
Prototype or workflow
Testable system, routing logic, data layer or automation block clean enough to review.
Action queue
Priorities, owners, controls and conditions for moving to the next level.
Contact
If the topic is real and priority-driven, the efficient move is to frame it quickly.
The first exchange is mainly there to confirm fit, recommend the right format and define the level of proof worth producing.