Insights
Practical thinking on the questions behind an implementation: what to use, where information goes, what it costs and how your team makes it work.
Start with your decision
These guides answer the questions buyers face before they commit data, budget or operating responsibility.
Pressure-test the plan
A six-pass review of data, integrations, failure handling, costs and operating ownership.
Read the guide →Choose an approach
A practical brief for comparing existing products, integration and custom implementation.
Read the guide →Make the economics work
A worked example covering model usage, review, hosting and maintenance.
Read the guide →Use proprietary information
Access, processing and deployment decisions before selecting a model.
Read the guide →Get ready to operate
Source mapping, ingestion, permissions and practical acceptance checks.
Read the guide →Own the result
Accounts, access, documentation, support and exit questions to settle before handover.
Read the guide →The library
Decision guides sit alongside dated industry commentary. Publication and update dates help distinguish the context of each article.
77 articles
AI Strategy & Transformation
A practical worksheet for comparing AI vendors against real tasks, data rules, integrations, full operating cost and exit work.
Published 2026-04-28 · Updated 2026-10-03
AI Strategy & Transformation
A decision worksheet for measuring complete AI workflow cost, reconciling usage with invoices and testing changes against accepted results.
Published 2026-05-02 · Updated 2026-09-30
AI Implementation Practical Guides
A practical launch test for source support, missing information, permissions, safe refusal and human review in business AI answers.
Published 2026-03-22 · Updated 2026-09-29
AI Implementation
A working AI-built prototype is useful evidence. Review its architecture, security, failures, costs and ownership before real users depend on it.
Published 2026-04-23 · Updated 2026-09-28
AI Data & Integration
Before an agent reads business data, prove that each delivery is complete, valid, traceable and recoverable when something goes wrong.
Published 2026-09-24
AI Implementation
Claude, ChatGPT or another AI assistant can produce a persuasive technical plan. The next step is proving that it fits your business.
Published 2026-09-23
AI Security & Risk
Decide what must remain private, which components touch the information and who will operate the environment before choosing an approach.
Published 2026-04-26 · Updated 2026-09-23
AI Implementation
A working demonstration is not a completed implementation. Verify that your company can access, operate, test, change and fund the system.
Published 2026-09-23
AI Implementation Practical Guides
Custom AI is justified by a specific unmet requirement. Test existing products first, then configure, integrate or build only the missing layer.
Published 2026-03-15 · Updated 2026-09-23
Technology Deep Dives
A practical checklist for connecting incoming files, databases and business systems to AI, including validation, access, freshness and recovery.
Published 2026-03-16 · Updated 2026-09-17
AI History
From the 2018 archives: Levi Brackman on consumer loyalty to AI agents, compounding intelligence and AI-to-AI travel distribution.
Published 2026-09-11
Technology Deep Dives
Evaluate local processing against quality, hardware, privacy requirements and operating costs.
Published 2026-05-05 · Updated 2026-09-11
Start a conversation
Tell us what you want to achieve and what needs to be true for it to work—your budget, your data requirements, your existing systems or your team’s capacity.
Discuss your AI projectWe’ll discuss whether we’re the right team to help and what a sensible next step could look like.