Connect the data and tools, set permissions, and test the workflow against real cases.
Internal AI Playbook / Book in progress
Build internal AI people can trust and use.
A field guide for engineers and operators building internal AI systems that need to work in production.
Get early chaptersWhat you will learn
Turn one AI use case into a system people keep using.
Production work starts with a business decision. Define the data behind it, the person who owns it, the actions AI may take, and the result you will measure.
The chapters apply this method to inbound, outbound, customer success, and expansion. The architecture and operating choices still matter when the models and tools change.
The system
A shared foundation for every revenue agent.
Shared foundationProduct knowledge + company context
Docs · CRM · usage · calls · tickets · policy
AI Outbound SDRAI Inbound SDRCustomer Success AIAI Expansion Agent
Control layerPermissions · approvals · evaluation · feedback
Who it is for
For the teams building and operating internal AI.
Turn routing, policy, and process decisions into workflows teams can run.
Design AI around customer context, frontline judgment, and measurable account outcomes.
Inside the book
Eight chapters, from first workflow to production.
- 01
Why go-to-market became an engineering problem
See the shift from disconnected tools to systems with owners, interfaces, and feedback.
- 02
The internal AI stack
Connect product knowledge, customer context, agents, tools, permissions, and evaluation.
- 03
Choosing the first production workflow
Turn a broad AI idea into one measurable decision with a clear operating boundary.
- 04
Product knowledge that agents can use
Build a maintained source of truth instead of sending every document to a model.
- 05
AI inbound and outbound SDR
Design research, qualification, routing, drafting, and review around pipeline quality.
- 06
Customer success and expansion AI
Use live customer context to surface risk, next actions, and credible growth moments.
- 07
Human control without workflow drag
Place approvals, permissions, limits, and escalation where consequence demands them.
- 08
Evaluation, adoption, and operating change
Measure model quality, workflow health, human use, and business outcomes together.
Written from practice
The work comes before the advice.
I build internal AI systems at an NYSE-listed company across product knowledge, inbound, outbound, customer success, expansion, and revenue automation.
The book turns that work into system blueprints, agent patterns, and production decisions you can adapt to your own company. Why I’m writing it.
Join the waitlist
Get the first chapters before launch.
Join for early chapters, working diagrams, and the launch announcement.