AI Scoping Lab
2–4 weeks of focused, hands-on technical work to answer "can we actually build this, and what would it cost?"
A scoping lab is for organisations that have already chosen a specific AI use case — from a workshop, a roadmap, or an internal idea — and now need to know whether it's technically feasible, what the data really looks like, and what a working solution would take.
Who it's for
- Teams with a specific use case in mind, not a portfolio of ideas
- Product, R&D, IT or innovation leads who need a credible build estimate before requesting budget
- Organisations preparing a Proof-of-Concept and want it scoped properly before the development team starts
What's included
- Use-case deep-dive: problem definition, success criteria, real-world constraints (latency, accuracy thresholds, integration points)
- Data readiness audit: what data exists, where it sits, what's missing, what cleaning or labelling is realistically needed
- Technical feasibility study: model choices (classical ML / fine-tuned / RAG / agentic / off-the-shelf), build-versus-buy analysis, key risks
- Architecture sketch: the working system in a diagram, with the components, the data flow, the integration points, the human-in-the-loop steps
- Proof-of-Concept plan: scope, milestones, evaluation criteria, exit criteria
- Effort and cost estimate: ranges for PoC, pilot and production, with stated assumptions
- AI Act risk classification and any GDPR considerations
- Final readout to your team plus a written report (~15 pages)
Duration: 2 weeks (focused single use case) to 4 weeks (use case with significant data unknowns or integration complexity).
Deliverable: enough technical clarity that you can decide, with confidence, whether to build, buy, defer, or kill the idea.
What this is not: the build itself. The Scoping Lab ends with a decision document, not a working model. If you'd like us to build the PoC afterwards, we can — but you're also free to take the plan to anyone else.