AI Blueprint · How we work together
The 4D methodology
I use Discover, Design, Develop and Deliver to guide our work together. We move back and forth as we learn, building and delivering value in small, useful increments.
Each phase informs the others. What we learn in use can reshape the opportunity, improve the design or guide the next increment.
01 / Discover
Understand your ambition, customers, current work and constraints. Establish what would make a meaningful difference.
Business strategy
A useful AI investment starts with the business you want to build.
We work through your goals, customer promise, strengths and commercial priorities in a focused leadership conversation. We identify where growth, margin, capacity or resilience needs attention.
You leave withA business context map and agreed priorities.
Customer research
The first request is not always the underlying need.
We speak with customers and frontline staff, examine recent enquiries and observe how people get work done. We separate evidence from assumptions and record questions still to answer.
You leave withA customer needs summary with evidence and open questions.
Customer journey mapping
Delays and repeated effort often sit between individual tasks.
Together we map a real journey from first enquiry to follow-up, using examples and customer conversations. We mark questions, friction, emotions and missed opportunities at each step.
You leave withA current journey map with priority friction points.
Process and handover mapping
Hidden coordination makes work difficult to improve.
We trace who does what, which information they need and where work waits. Staff walk us through real cases, including exceptions and the informal workarounds that keep things running.
You leave withA current process map with handovers, delays and owners.
Technology and information review
Existing systems, knowledge and data shape what is practical.
We review the tools you use, available integrations, information quality, permissions and operating constraints. We identify useful existing capability and the gaps that could block delivery.
You leave withA systems and information inventory with constraints and gaps.
Opportunity assessment
Not every possible automation deserves investment.
We compare opportunities by customer value, business impact, effort, uncertainty and risk. Together we define a baseline, success measures and the assumptions a small experiment should test.
You leave withA prioritised opportunity backlog and measurable experiment brief.
Before moving on
A shared opportunity brief and a decision about what deserves exploration.
We review the evidence together and agree whether to proceed, adjust or stop.
02 / Design
Shape the experience, workflows and technology. Test important assumptions before committing to a larger build.
Future customer journeys
A faster task should contribute to a better overall experience.
We sketch the future journey with customers and staff, showing what changes and what should stay human. We test the sequence against realistic situations and the outcomes agreed in discovery.
You leave withA future journey map showing the intended experience.
Service blueprinting
A good experience depends on more than the visible interface.
We align customer actions with staff work, agent activity, supporting systems and information flows. We make dependencies and responsibility clear across the visible and behind-the-scenes service.
You leave withA service blueprint linking customer experience to delivery.
Workflow and agent design
Useful agents need boundaries, context and a way to act.
We define triggers, information sources, tools, decisions and handovers. We design when an agent proceeds, when someone reviews its work and how exceptions or uncertainty are handled.
You leave withAn agent and workflow specification with review points.
Experience design and prototyping
A tangible example reveals misunderstandings early.
We make sketches, interface prototypes or a small working demonstration. People try representative tasks while we observe what they understand, where they hesitate and what needs to change.
You leave withA tested prototype with findings and design decisions.
Solution architecture
Reliable delivery depends on the joins between components.
We map applications, models, data, integrations and hosting. We work through access controls, failure handling and ownership, and compare build, buy and reuse options against your constraints.
You leave withA solution architecture and a record of key decisions.
Technology strategy and roadmap
A clear direction helps avoid disconnected tools and unfinished pilots.
We agree the technical direction, dependencies and delivery increments. Each increment has a useful outcome, an owner and a decision point; costs, subscriptions and support assumptions are made explicit.
You leave withA technology roadmap with scope, dependencies and indicative investment.
Before moving on
A tested direction, a solution blueprint and a practical delivery roadmap.
We review the evidence together and agree whether to proceed, adjust or stop.
03 / Develop
Build and connect the solution in useful increments, with your team involved throughout.
Pilot development
A bounded pilot exposes practical constraints before wider investment.
We build the smallest useful slice and agree who will try it, with which cases and under what limits. Regular demonstrations keep the work grounded in the original business need.
You leave withA working pilot and a record of what it proves.
Application and agent development
Good ideas need usable interfaces and dependable behaviour.
I build applications, agents and workflow components in reviewable increments. We define expected behaviour, work through real tasks and refine the experience with your team as it takes shape.
You leave withWorking software with source, configuration and agreed ownership.
System integration
An answer becomes more useful when it can move the work forward.
We connect agreed systems through suitable APIs, events or other supported interfaces. We test permissions, duplicate requests, failures and recovery so handovers behave predictably.
You leave withTested integrations and documented information flows.
Knowledge and data preparation
An agent cannot consistently help with information it cannot trust or access.
We organise relevant business information, resolve gaps and define access. We test retrieval against real questions and agree how content will be updated and who is responsible for it.
You leave withA maintained knowledge structure with access and update rules.
Quality, safety and reliability
A convincing demonstration is not enough evidence for everyday use.
We test representative tasks, edge cases and failure paths. We check answer quality, security, accessibility, performance and cost against agreed criteria, recording limitations and remediation work.
You leave withAn evaluation set, test results and a readiness assessment.
Feedback and iteration
Real use will reveal things the initial design missed.
We review pilot feedback and performance with users, prioritise changes and retest. Findings may send us back to discovery or design rather than simply adding more features.
You leave withAn improved release and a prioritised backlog of next changes.
Before moving on
Working capability, evidence of its performance and a decision about readiness.
We review the evidence together and agree whether to proceed, adjust or stop.
04 / Deliver
Introduce it safely, support the people using it and measure whether it delivers the intended results.
Rollout and adoption
A working system only creates value when it fits everyday work.
We plan a staged rollout with clear responsibilities, communications and a fallback. We agree what must be true before expanding use and give the team a way to raise problems early.
You leave withA rollout plan, readiness checklist and adoption actions.
Training and enablement
People need to understand both the capability and its limits.
We practise real tasks with the team, explain when to question or override outputs and create concise guidance. We identify local champions and the support people need as habits change.
You leave withPractical training, task guides and escalation guidance.
Ownership and handover
Ongoing operation should not depend on undocumented knowledge.
We agree who owns the service, source, accounts, data and decisions. I document how it runs and arrange the agreed handover, maintenance and support responsibilities with your team.
You leave withA handover pack, operating guide and responsibility map.
Monitoring and operations
Models, information and connected services can change over time.
We put appropriate monitoring in place for failures, quality, usage and cost. We define alerts, review routines and recovery steps, with named people responsible for responding.
You leave withAn operational dashboard and an incident response runbook.
Outcome measurement
Usage alone does not tell you whether the investment was worthwhile.
We compare results with the baseline and gather feedback from customers and staff. We examine saved effort, commercial outcomes and quality, including where benefits have not appeared.
You leave withAn outcome review with evidence and recommended decisions.
Continuous improvement
The best next step may be refinement, expansion or stopping.
We review what we have learned, revisit business priorities and choose the next useful increment. New opportunities return to discovery with evidence from the work already delivered.
You leave withAn updated improvement backlog and next-step roadmap.
Before moving on
An adopted service with clear ownership, support and an improvement plan.
We review the evidence together and agree whether to proceed, adjust or stop.
We agree a useful outcome and how it will be measured, with a baseline and clear numerical targets for success. We agree a fixed fee and how we’ll demonstrate that those targets have been met.
If I don’t deliver the agreed result, you don’t pay the delivery fee.
Start with a useful conversation
Bring an ambition, a recurring problem or something you have already started. Together, we’ll identify what is worth doing next.
Arrange a discovery callFree · 30 minutes with Martyn · No obligation
Martyn AIChat about your possibilities