When strategy is the right starting point
You may have several AI ideas but no agreed priority, or a pilot that has not moved into everyday use. A discovery engagement helps separate a promising use case from an interesting demonstration. We look at the business process, the people involved and the evidence needed to make a decision.
- Teams repeat research, reporting or document tasks.
- Leadership needs a realistic scope and cost model.
- Existing experiments lack ownership or success criteria.
What the engagement includes
We map candidate workflows with the people who do the work. We then review data availability, language requirements, integrations and operational constraints. Each opportunity is assessed for value, feasibility and risk so the roadmap reflects your situation.
- A prioritised opportunity shortlist with reasons.
- A readiness review covering data, access and ownership.
- A pilot brief with boundaries, evaluation criteria and dependencies.
- A delivery roadmap and assumptions behind the estimate.
A practical business example
Consider a professional services team that prepares proposals using Arabic and English reference material. The first step could be an assistant that finds approved source passages and drafts a section for review. It should not start by automatically sending proposals to clients. This is an illustrative scenario, not a Momkin client result.
- Measure time to prepare a reviewed draft, not only generation speed.
- Test both languages and preserve source references.
- Keep commercial commitments with the responsible team.
How we decide whether to proceed
A pilot should have an explicit decision at the end: proceed, revise or stop. We define a baseline, a representative test set and the level of review the workflow needs. If a simpler automation solves the problem, that should be part of the recommendation.
- Name the person who owns the outcome.
- Include integration, support and review effort in the cost.
- Agree what evidence would justify the next investment.
Before we begin.
What should we prepare for a discovery conversation?
Bring one or two frustrating workflows, a description of the systems involved and non-sensitive examples. You do not need a technical specification.
How long does an AI strategy engagement take?
Duration depends on the number of workflows, stakeholder availability and access to representative information. We agree the scope and milestones before work begins.
Do we need to commit to a full implementation?
Discovery can be scoped as a separate decision stage. Its output should help you decide whether implementation is worthwhile.