"We have lots of data but do not know what to do with it." We hear this in almost every SME meeting. Here is our complete method to go from this situation to a first concrete deliverable in 6 weeks.
Why most SME data projects fail
The pattern is always the same: leadership decides to 'do data', buys a BI tool, asks IT to configure it, and 6 months later nobody consults the dashboards. Cost: €20,000-50,000. Result: zero.
The fundamental mistake: starting with the tool rather than the problem. Technology does not create value by itself. Decisions made thanks to data create value.
Our 4-phase method
Phase 1: mapping (weeks 1-2)
Interviews with decision-makers and operational teams. Objective: identify the 10 key business decisions and understand what data is available.
Phase 2: prioritisation (week 3)
Rank use cases by impact and feasibility. Retain the 3 use cases in the high impact, high feasibility quadrant.
Phase 3: quick wins (weeks 4-5)
Deliver a first concrete result quickly on the simplest use case. Not perfect, but real and measurable.
Our conviction: an SME does not need a Data Lake or a dedicated data team to start. It needs 3 well-defined indicators, consulted weekly by the right people.
With care,
Excellent article, this matches exactly what we're seeing with our enterprise clients. The section on inference costs is especially valuable. It's a topic most articles gloss over but it's make-or-break at scale.
Thanks James! Inference cost optimization is often deprioritized during prototyping but becomes critical in production. Feel free to book a session if you'd like to go deeper on this.
Sharing this with my whole team. The distinction between an impressive demo and robust production is exactly the debate we're having internally right now. The human checkpoint advice is immediately actionable.
Great article. I'd push back slightly on the 18-day deployment estimate, in our experience with enterprise security and GDPR requirements, 4–6 weeks is more realistic for a first production agent.
Completely fair point David. The 18 days refers to a scoped first agent in a test environment. For full enterprise production with security constraints, your estimate is accurate.