AI strategy and discovery
Every idea in the business, captured, scored and ranked by the people who do the work
Divergent workshops to open the field as wide as it will go, then convergent workshops to bring it under control. The output is not a list of ideas; it is a prioritised, traceable set of opportunities with the data position and readiness of each.
Open wide, then bring it under control
Most AI idea-gathering fails in one of two ways: it filters too early, so the people who know where the pain is never raise it, or it never filters at all, so the business is left with a hundred sticky notes and no decision. The divergent-convergent approach is built to avoid both.
Divergent sessions are organised by function, bringing together the people who do the work alongside colleagues from product management and IT. Their purpose is breadth, not judgement: every idea for where AI might help, however large or small, logged against the business area, the pain or opportunity, how it is handled today, and the AI capability it would call on. Nothing is filtered for feasibility at the point of capture.
Convergent sessions revisit each function's output. Ahead of the workshop we pre-cluster the ideas into related groups and pre-score them on the ten elements of the convergence scorecard, deliberately as a straw man for the group to challenge, correct and force-rank. The scores that come out are the business's own view, calibrated in the room, which is what makes the resulting roadmap one the organisation acts on.
10Workshops in our last full programme, five divergent and five convergent
250Ideas captured across five functions
53Clusters scored and force-ranked by the people who raised them
A full programme across five functions
A smaller sprint covers one or two functions in four workshops.
Preparation
Functions and participants agreed, the capture template adapted to your terms, and the awareness-session vocabulary carried in so everyone starts from the same picture.
Divergent workshops, one per function
Two to three hours each. Facilitated, not presented. Every idea captured with its pain, current process, proposed AI use, benefits and beneficiaries, and tagged with its layer on the Opportunity Pyramid.
Clustering and pre-scoring
Between sessions we cluster each function's ideas, score every cluster on the ten elements, and identify the common themes that recur across functions, which are usually the strongest opportunities the business has.
Convergent workshops, one per function
The straw-man scores challenged and corrected, clusters force- ranked, and the cross-business themes shared with every group so each can see what it has in common with the others.
Consolidation
Ideas sharing a capability are consolidated into submittable use cases, each assigned an adoption pathway and a data position, and sequenced by readiness. See opportunity prioritisation.
A complete, traceable record
The discovery workbook: every idea with its full capture, its cluster, and the convergence scoring and comments behind the priorities, so that any decision in the final roadmap can be traced back to the room it came from. The scored cluster views per function. The cross-business common themes. The appetite-versus-readiness view that shows where enthusiasm and data pull in opposite directions. And the consolidated presentation, with every figure editable, for use inside the business afterwards.
In practice, discovery is run with opportunity prioritisation as one programme, and the findings report is the combined output.
Four weeks to twelve
A sprint across two functions runs four workshops over about four weeks. A full programme across five functions runs ten workshops over eight to twelve weeks, with the findings report and readiness-sequenced roadmap at the end. Scoped and fixed before it starts.
About discovery workshops
How many people take part?
Six to fifteen per workshop, with product management and IT represented inside each function rather than treated separately. Our last full programme involved close to seventy people.
Which functions should we include?
The ones where the work happens: in a manufacturer, R&D, manufacturing, procurement and supply chain, sales and service, and marketing. Finance and legal are often under-represented and worth a dedicated short session.
How long between the two stages?
One to two weeks, which is the time we need to cluster and pre-score each function's ideas before its convergent session.
What if the ideas are all about the same few things?
They usually are, and that is the finding: the same request for an assistant, the same need to search scattered knowledge, the same appetite to forecast, raised independently in every function. Those become the cross-business themes, funded once and returning value many times over.
Do you run the workshops or do we?
We facilitate; your people decide. The pre-scoring is ours, the final scores are theirs. The published artefacts let you run later rounds yourselves if you want to.
Find out what your people already know
The ideas are in the business. Discovery is how they become decisions.