Architecture
Describe the whole system.
Components, interfaces, hosting, data flows and operational responsibilities, at the level needed for a sensible implementation.
Blueprints and proof · from £1,950
Make the next decision credible.
Good ideas become expensive when nobody has described the system, the risks or the evidence needed to make them work.
Solution Blueprint
Know what to build or buy.
A vendor-neutral design your internal IT team, implementation partner or procurement team can actually use.
The Blueprint takes a worthwhile opportunity and gives it enough structure to move forward: functional and non-functional requirements, architecture, data flows, authentication, security boundaries, costs, dependencies and acceptance criteria.
Judgement
Choose proportionately.
Build versus buy, hosted versus private, conventional software versus AI, with the trade-offs made explicit.
Handover
Leave with something useful.
A design you can take to internal IT, an MSP, a development partner or procurement without starting the conversation again.
Blueprint contents.
| Problem and assumptions | What is being solved, what is known, and which uncertainties still matter. |
| Requirements | Functional, non-functional, operational and acceptance requirements. |
| Technical design | Major components, interfaces, identity, authorisation, data and hosting. |
| Safety and resilience | Security boundaries, privacy considerations, failure modes, testing and recovery. |
| Delivery route | Phases, dependencies, indicative costs and questions for potential suppliers. |
Prototype the dangerous assumption.
From £2,950, a Prototype Sprint answers one fixed question against a fixed dataset in a fixed timeframe. Can the API integrate? Can OCR recover the fields? Can search find the right report? Can a local model perform well enough?
Question
One assumption that could make or break the project.
Evidence
Agreed sample data and a measurable definition of success.
Test
A deliberately small implementation, not a disguised production build.
Verdict
Proceed, change direction or stop before the expensive commitment.
A failed prototype can be a successful engagement. “This does not work reliably enough” may be the conclusion that prevents a £50,000 mistake.
Start with the question that matters.
Most Blueprint work follows an Opportunity Review, but a well-defined technical question can start here.