01 / Index
Shipping a public LLM endpoint without a surprise bill
how to add an llm feature without runaway api costs
The cost controls on a public, unauthenticated AI endpoint in production: per-IP rate limiting, structured outputs, low reasoning effort, and a static fallback that keeps the feature working when every one of them trips.
5 minWhat a $3,000 AI pilot actually buys you
how much does it cost to build a custom AI tool for my business
A line-by-line breakdown of what a fixed-price $3,000–$6,000, two-to-three-week AI pilot includes, what it deliberately excludes, and why fixed price beats hourly for a first engagement.
6 minBuild vs. buy for an internal AI tool, at 5–50 people
should we build or buy an internal AI tool
When a SaaS subscription is the right answer for a small B2B company and when it isn't, framed around switching costs, per-seat pricing, and who owns the system at the end.
6 minScoring B2B leads with an LLM, and where it fell over
how to score b2b leads with an llm
A teardown of the ANTA Lead Intelligence Agent: why the off-the-shelf lead scores were unusable, what replaced them, and the three failure modes that showed up once it was scoring real accounts.
6 minSeven questions to ask an AI agency before you sign
what to ask an ai agency before signing a contract
The questions that separate a studio that will ship you a working system from one that will bill you for a discovery phase: repo ownership, model cost exposure, handover, and what happens when the pilot fails.
5 min
You already know what to build.
Send the two-paragraph version. You'll get a real technical response, not a calendar link and a discovery questionnaire.