Ending 01
Build it
The assumption held. The work continues as an implementation, with the architecture already decided and tested.
Implementation →Engagement · bounded
One concrete Data or AI problem, scoped and moved forward by a senior partner — without opening a large consulting programme. The output is a decision you can act on, plus the technical work that supports it.
“We have a concrete Data or AI problem. Can a senior external partner scope it and move it forward without starting a large consulting programme?”
Use cases
Select a case to see what the engagement would actually do with it.
A difficult analytics problem
What the Sprint does
Reconstruct the question, find where the numbers diverge, and rebuild the model or metric definition so the answer holds up under review.
Typical output
A working analysis, a defensible definition, and the pipeline or query layer behind it.
How it runs
01
What decision depends on this, who needs it, and what would count as a good answer.
02
Data available, technical approach, constraints, and the boundary of what the Sprint will cover.
03
A prototype, model or pipeline built far enough to test the assumption the decision rests on.
04
A recommendation with its limits stated, the artefacts, and documentation your team keeps.
Duration and cost follow the problem. Not every project fits a fixed length or a fixed price, and a Sprint is not sold as one.
// Where a Sprint ends
A Sprint is not a sales step towards a larger programme. Two of these three endings are cheaper for you than the third.
Ending 01
The assumption held. The work continues as an implementation, with the architecture already decided and tested.
Implementation →Ending 02
The data, the economics or the problem definition does not support it. You get the reasoning, the evidence, and what would need to change first.
Ending 03
The Sprint reveals that this problem is one of several, and the gap is senior direction across all of them.
Fractional Lead →// Comparable work
SEC 01 / PIACRADAR / SYNTHETIC DATA · REPOSITORY FACT
LLM market-intelligence workflow. One question becomes hypotheses, evidence, and an action plan.
Sprint pattern: a strategic question is turned into a structured, reviewable decision package rather than a one-shot answer.
Inspect the case study →SEC 02 / TASTETREND / DEMO AVAILABLE
AI RAG over restaurant reviews. Ask a question, inspect the evidence, open the demo.
Sprint pattern: a retrieval approach built far enough to judge answer quality and evidence handling before committing to production.
Open the work index →Describe the situation in a sentence or two. If a Sprint is the wrong shape for it, that will be clear in the first conversation.