- Use-case shortlist with estimatesEach candidate with a build estimate, a running-cost estimate and an expected return, with the assumptions written next to the figures so you can change them.
- Proof-of-concept with sourceA working prototype of the strongest candidate, delivered with its source and notes on what a production version would need.
- Go/no-go recommendationA written recommendation on whether to proceed, what the build would cost, and what would have to be true for a no to become a yes.
Services/AI Consulting & Proof-of-Concept
AI Consulting & Proof-of-Concept
Find the use case worth building, then prove it in two weeks.
What it is
Teams new to AI rarely lack ideas; they lack a way to tell which idea deserves the first budget. This engagement starts with a working session on your actual processes and data, narrows the ideas to a shortlist of candidate use cases, and attaches to each an estimate of what it would cost to build and run and what it would return.
Then one candidate is built as a small proof-of-concept, with source, that real users can try before you commit to a full build. It ends with a written go/no-go recommendation. A recommendation not to proceed is a real outcome of this work, and a cheaper one than finding out after the build.
What you get
How the work runs
Working session
A structured session with the people who own the process: what is repetitive, what touches data, where mistakes cost money. The output is a long list of candidates and the facts needed to estimate them.
Shortlist with cost and return estimates
Candidates are scored on feasibility and value. Each one that survives gets an estimate of build cost, running cost (model usage, hosting) and return, with its assumptions stated.
Two-week proof-of-concept
The strongest candidate is built as a working prototype with source, using tools a production build could keep, and put in front of real users.
Go/no-go
A written recommendation: what users did with the prototype, what it cost to run, what a production version would take, and whether to proceed. If yes, the prototype and the estimates become the brief for the build.
Pricing
Consulting
$100–$200/hrUse-case workshops, estimates and review.
Proof-of-concept
$1,000–$3,000A working prototype with source and a go/no-go recommendation.
Prices are in USD and are starting points. Consulting is billed by the hour and the proof-of-concept is a fixed price; the final quote follows a scoping call.
Where it fits
A good fit when
- You run a startup or small business and know AI should fit somewhere, but not where.
- You have been quoted a large build and want evidence before you commit to it.
- You need something real to show users, a co-founder or an investor, not a slide.
Not a fit when
- You already know the use case and want it built. Go straight to Conversational AI & RAG Agents or Agent Orchestration & Workflow Automation.
- You want a long strategy document. The outputs here are short: a shortlist with estimates, a prototype and a recommendation.
- You need an existing system measured rather than a new one tried. See Agent Evaluation & Benchmarking.
Questions
What if the recommendation is no-go?
You keep the shortlist, the estimates and the prototype source, and you have spent the proof-of-concept budget instead of the build budget. The recommendation says what would have to change for the answer to become yes.
Who owns the proof-of-concept?
You do. It is delivered with source and notes on what a production version would need, whoever builds it.
Can I book consulting without a proof-of-concept?
Yes. Consulting is billed by the hour and covers use-case workshops, estimates and review. The proof-of-concept is a separate fixed-price piece of work.
How are the estimates produced?
From your numbers: how often the task happens, how long it takes today, and the published prices of the models and hosting involved. Every figure has its assumptions written beside it.
Find out before you build
Describe your business and the work that takes the most time. The first step is a working session; the last is a recommendation you can act on.