10 articles
Buying and budgeting AI
Prices, contracts and build-or-buy calls, with the numbers written down. What an agent actually costs to run, what a proposal should contain, and what you own when the work is finished.
-
Measuring cost per resolved task
Total AI spend tells you nothing. The unit metric that does, how to compute it including the human cost of escalation, and why it beats deflection rate.
-
Fractional AI lead: when it beats hiring
What a fractional AI lead actually does, when it is the right shape against hiring or using an agency, and the four conditions under which it fails.
-
Pilot to production: what the second invoice covers
Why the gap between a working AI pilot and a production system is usually two to four times the pilot cost, and what that money actually buys.
-
Cutting LLM costs without degrading quality
Eight cost reductions ranked by quality risk, from the ones that are free to the ones that are a real trade - and how to prove which is which before shipping.
-
Shopify AI app or custom integration
When an app from the store does the job, when it hits a ceiling, and the three questions that decide it before you spend anything.
-
What you own after an AI project: code, prompts and evals
The eleven artefacts that decide whether you own an AI system or rent it, the contract language that secures them, and the handover test that proves it.
-
Fixed price or time and materials for AI projects
When a fixed price protects you on an AI build, when it quietly costs you more, and the phased structure that avoids the worst of both.
-
Build or buy AI customer support
A decision framework for AI support: five conditions that make buying correct, four that make building correct, and the hybrid most companies should actually run.
-
How to evaluate an AI agency proposal
Fourteen questions that separate an AI proposal that will ship from one that will produce a demo, plus the answers that should end the conversation.
-
Scoping an AI project that actually ships
Most stalled AI projects were scoped wrong on day one. The predictor is whether the scope was drawn around a workflow or a capability, and four questions settle it.
The other folders
- Which tool, and why do these things break? Workflow automation 13 articles
- How should the system be put together? Building and running it 11 articles
- How do we know it is any good? Measuring whether it works 11 articles
- What do we have to be able to show? Governance and regulation 4 articles
- Why does ChatGPT name a competitor? AI search visibility 4 articles
Let's talk
Prefer a conversation to an article?
Most of what is written here started as a question a client asked on a call. If you have one, bring it - we answer questions before contracts.