Why Enterprise AI Stalls After the Demo Start here
Why pilots that impress in a demonstration stall in daily operation, and what changes that. The short one to forward to a colleague.
Whitepapers
Nineteen whitepapers on getting AI into production.
Free to download, no form. Written for the business owner, the architect and the risk reviewer who each have to say yes.
For executives and program owners
Why AI stalls, what to fund, and what an engagement looks like.
Choosing an Enterprise Platform for AI Agents
How to choose ChatGPT, Claude, Copilot or Gemini; the right models; Bedrock, Foundry or Google; and the agent platform, with hard gates, a scorecard and a five-week proof plan.
The Connected Enterprise Data & AI Blueprint
The executive guide: the architecture and operating decisions that move enterprise AI into production.
Adoption by Design: The Change Management That Decides Whether Enterprise AI Is Used
Why working AI still goes unused: owners, hubs and hackathons, the public evidence, and an adoption worksheet for the business owner.
The 90-Day First Governed Capability
A thirteen-week plan for the first governed capability, with three gates and the point where Build continues.
For data leaders building the operating model
The 2026 practitioner series: ownership, products, platform and governance.
The Data and AI Operating Model
How to organize ownership, delivery teams, decision rights and funding, from a first team to the default way of working.
Data Products in Practice
Product boundaries, types, contracts, master data, quality and certification: what domain teams build and own.
The Federated Data Platform
Shared services, delivery paths, semantics, copy or query, and agent access, built on the stack you already run.
Context Before Query
Give agents business meaning before the call and keep policy enforced at the data, for tables and documents alike.
For architects and engineers
The target state, the products, and the contracts your teams will build.
The Four-Part System
The reference architecture: four responsibilities, the brain, nervous system, body and immune system, and the control plane that links them.
The Reference Implementation: AI, Data and Security as One System
The full picture with real products: layers, labels and tags, agent identity and supervision, the security operations center, and a bank and an insurer traced end to end.
Patterns and Product Fit: Securing AI Retrieval at Enterprise and Lean Scale
The engineer’s guide: Cyera, Purview, Snowflake, Databricks and data.world working together, permission-aware retrieval, SharePoint, vector stores, AtScale, and two lean stacks.
Agent Onboarding: Treating an Agent Like a Data Product
Treat an agent like a data product: tool contracts, identity, evaluation and continuing ownership.
Data Product Certification Standard
A practical standard for certifying a data product: meaning, access, quality and ownership.
For risk, compliance and audit reviewers
What a reviewer will ask for, and a worksheet for the three owners.
AI Governance Evidence
What reviewers inspect for AI at US banks and insurers: SR 26-2, Regulation B, NAIC and NYDFS rules, request traces and release records, plus EU and UK rules for international firms.
The 10-Question AI Production-Readiness Diagnostic
Ten questions for the business owner, technical lead and control owner to answer together before a workflow goes live.
Operators’ Corner
Practical guides and tools for the people who run AI and data day to day.
AI Token Economics
For the CFO, CIO and CTO: forecast AI spend by workflow, separate capacity from cash, and decide with allocated, avoidable and incremental views. With an Excel forecast model.
Data Is Not Just Code
Why connecting a model to data does not produce reliable answers on its own: context, authority, evidence, and a release pipeline that matches the consequences.
Rewiring Data Management with AI
For data leaders and engineers: approved contracts and bounded agents across the data lifecycle, with checked economics, a claims example and templates for a first workflow.
Next step
Bring one workflow and the decision in front of you.
We will work through what is blocking it, what evidence is missing and whether there is a useful next step.