AI Product Consulting
Strategy before code. Model choice, architecture, cost at real volume, and an honest read on whether the thing should be built at all.
Strategy before a line of code.
Model selection, architecture review, cost modelling, and an honest roadmap your own team can execute. Sometimes the most valuable output is talking you out of a build.
You probably need this if
Three vendors have quoted wildly different numbers and you cannot tell why.
There is a build underway and nobody can say what it will cost at scale.
Your team is strong but has not shipped an LLM system before.
The board wants an AI strategy and you want one that survives contact with reality.
What the engagement includes
Architecture review
What you have, or intend to build, assessed for the places it will hurt: latency, cost, evaluation, and the parts that quietly do not scale.
Model and vendor selection
Benchmarked on your task with your data, rather than read off a leaderboard.
Cost modelling
Per request, per user, per month, at the volume you actually expect, including the retries nobody budgets for.
A roadmap your team can run
Sequenced, sized, and written for the people who will build it rather than for the people who approve it.
What you end up holding
- A written assessment with the trade-offs made explicit
- Benchmarks on your own data
- A cost model you can put in front of a finance director
- A roadmap, and a second opinion whenever you want one
Asked often enough to answer here
- Will you tell us not to build it?
- Sometimes, and it is usually the most valuable thing we say. A month of consulting that prevents a wasted year is a good trade.
- Do you only consult on work you would build yourselves?
- No. Plenty of engagements end with your own team building it, which is a perfectly good outcome.
- How long does an engagement run?
- Two to four weeks for a review and a roadmap. Longer if you want us alongside the build as a second pair of eyes.
Tell us what the work looks like now.
A short conversation is usually enough to say whether this is a week of work or a quarter, and what it would cost.