Delivering AI
You've been asked to deliver AI
You understand the theory, maybe from a certificate course. You need to see what a real solution looks like on your systems, and who you need to build it.
→ Start with the playbooksArchitect, engineer and technical project manager in one senior person. 20+ years building the ERP, warehouse, EDI, CRM and data systems that AI now has to plug into. AWS, Azure or your own stack. Honest about what you need, including the people I'm not.
Overview
Coming soon. Until it's ready, the playbooks below show the same thing in writing.
Delivering AI
You understand the theory, maybe from a certificate course. You need to see what a real solution looks like on your systems, and who you need to build it.
→ Start with the playbooksHave a partner
The strategy is agreed. Now it has to work with your core systems, and pass security and compliance.
→ How I work alongside your partnerExisting client
The board wants AI. I already know your systems, your data and your people, so adding AI is a continuation, not a new start.
→ How I'd approach itConsultancy or SI
You've won the programme and need a senior architect-engineer who can work on your client's AWS or Azure estate and legacy systems, under your governance.
→ How I work inside your teamOption A
“I'll take responsibility for the design, and make sure it works on your systems.”
Option B
“Your architect stays in charge. I make their job easier and their design stronger.”
Six roles, one person
Each one is a named gap in your RACI, not a parallel workstream.
Owns the target architecture across AI, data, middleware and your systems of record, and signs off design decisions.
Turns the programme plan into engineering work, runs technical risks and phase gates, and translates between the board and the engineers.
Builds the hard, hands-on pieces: APIs, events, pipelines and infrastructure as code, on AWS, Azure or your own stack.
Retrieval, agents, document AI and predictive models, tested and monitored on your own data.
Threat modelling, DPIA support, AI-specific threats, UK GDPR and PCI DSS, from the first gate.
A second pair of eyes on a partner's or vendor's design before you commit to it.
The playbook library
Worked examples: the business problem, the UK rules, the architecture on AWS, Azure or your own stack, the agents, the middleware, the delivery plan and the team. All 80 of the 80 planned playbooks are published.
No playbooks match those filters yet.
Try fewer filters, or tell me what you're working on. If it isn't here, it's probably a variation of one that is.
Patterns
Every playbook is assembled from these eight patterns. Learn them once, recognise them everywhere.
Read forms, letters and PDFs into structured data.
Answer questions from your own documents, with sources.
AI that takes small, controlled actions across your systems.
Put people on the decisions that matter, with the evidence.
Score, rank and forecast from your own historical data.
Connect AI to core systems safely, through events and queues.
Inspect photos, video and camera feeds for defects and damage.
Turn calls and conversations into text, notes and actions.
The same capability on each platform, row by row, with UK regions.
Open →UK regulation mapCross-sector rules first, then what each sector's regulators expect of AI.
Open →UK horizon scanWhat's coming from 2026 to 2030, and which playbooks it affects.
Open →Sovereign cloud, explainedA neutral guide: what it is, why some organisations care, the options, and how to decide.
Open →Delivering an AI programmePhases, gates, RACI and the business case, from course theory to real projects.
Open →Working alongside your partnerWhere I fit with a consultancy or systems integrator, and for existing clients.
Open →For students
Use these playbooks as worked examples. Each one has the business case, measures of success, risks and RACI your coursework asks for. I hold a Leadership with AI certificate myself, so I know what the courses cover, and where real projects differ.
I'm one person, not a 200-person consultancy. Most AI programmes don't need 200 people. They need one senior person who understands the business, the systems and the technology, and who can tell the rest of the team what to build. Where the work needs more hands, I bring in people I trust, or work inside your partner's team.
No. Each playbook is an illustrative scenario with typical numbers, written to show how a real solution fits together. Where one draws on my own work, it links to the anonymised case study and says so.
Whichever you already run. Every playbook shows the same design on AWS, on Azure, and on an open, portable stack you can run in your own data centre or with a UK or European provider. On Google Cloud? The same patterns apply. Ask me.
Either, and both are real offers. If there's no architect, I'll lead the design. If you already have one, internal or from a partner, I work under their lead and make their design stronger. We agree which on day one.
Most AI programmes need one senior person who can join up the business, the systems and the technology. Where the work needs more hands, I bring in people I trust, or work inside your partner's team.
Yes. They're written as worked examples, with the business case, measures, risks, RACI and delivery plan your coursework or programme board will ask for. A link back is appreciated.
What you're delivering, on which systems, and who's already on the team. I'll tell you honestly where I'd fit, and where you need someone else.