The AI Play Book is JMAN's reference for scoping, designing, and delivering AI solutions, from the concepts underneath them to the platforms they run on. It is written for technical and commercial teams alike, so an engineer and a consultant work from the same definitions, patterns, and trade-offs.

The Playbook at a Glance

Four sections, read left to right: why AI, what it is made of, how the parts combine, and where you build it. Each circle opens that page.

Fundamentals
Why and what
Core Concepts
The parts
Solution Patterns
How they combine

Pick Your Path

The same playbook, three routes through it. Start with the one that matches why you opened it.

Follow One Use Case

One question runs through the playbook: a portfolio company's finance team asking "What was revenue last quarter by region?" in plain English, and expecting a governed answer. Follow it from the concepts it depends on to the platform that answers it. Every page on the route carries a Running example: NLQ box.

Core Concepts
What it depends on
Solution Patterns
How it is built
Platforms
Where it runs

I Need To...

I need to...Go to
Explain AI to a clientFundamentals
Choose a platformPlatform Comparison
Estimate what it will costTokens, then the cost model on each platform page
Answer questions from documentsRAG
Let people query data in plain EnglishText-to-SQL and The AI Layer
Make it safe to put in front of usersAI Security, Safety & Governance

In This Release

This first release covers the foundations every later section will assume. Outlined squares are planned for later releases.

Coming Next
Planned
Use-case Scoping
Delivery Process
Standards
Evalu­ation
Case Studies
Gover­nance

Next: Fundamentals - start with what AI, machine learning, and generative AI actually are, and what they are not.