AI Opportunity Assessment for Small Businesses and SaaS
A fixed-price, written review of where an LLM or vision model would really help your business or product, and where it shouldn’t be used. Each idea gets a feasibility, risk and cost view, plus a ranked shortlist. It is for small businesses and SaaS teams. Prices start from $480 ($360 founding), and Starter takes about a week.
See a sample report, proposal and weekly update- Price
- from $480
- Founding
- from $360
- Timeline
- ~1 week
- Format
- Fixed scope, fixed price
- Hours
- US Eastern Time, async
Who this is for.
A good fit if
- Owners with repetitive text or image work: listings, catalog copy, document intake, support triage.
- SaaS teams under pressure to “add AI” who want one feature that users will actually use.
- Teams whose AI pilot works in demos but not on real data.
Common triggers
- Staff keep rewriting the same kind of text or data by hand.
- A competitor launched an AI feature.
- An AI prototype makes things up and nobody trusts it.
- You want to use AI but worry about customer data leaving the business.
Not a fit if
- You want a company-wide AI strategy, training program or change plan.
- You need self-hosted or fine-tuned models. I work with hosted model APIs.
- You expect guaranteed savings or accuracy numbers before any testing.
What I do, and what you get.
What I do
- Review your workflows or product and list the candidate AI use cases.
- Score each one on value, data readiness, error risk and running cost.
- Test the top candidates on a handful of your real, anonymized examples.
- Define guardrails: what the model may do, what must come from your data, and where a human reviews.
- Estimate model and API costs per month at your volume.
- Recommend a first use case, with a build estimate.
What you get
- A ranked shortlist of use cases in a value-vs-feasibility table.
- For the top pick: inputs, outputs, guardrails, review step, cost estimate and build estimate.
- Sample outputs from the test examples, with where the model failed.
- A “don’t use AI for this” list.
- Plus plan: a working prototype of one use case on your sample data.
- Handover notes.
Your part
- 10–30 real examples (anonymized if needed) of the work in question.
- A person who knows the workflow and can judge output quality.
- Your own model provider account if the prototype uses paid APIs (you pay the token costs).
Included
- Use-case list and scoring
- Small test on your real examples
- Guardrail and review-step design
- Monthly model-cost estimate
- Prototype (Plus only)
Not included
- Company-wide AI strategy or training
- Guaranteed accuracy, savings or ROI
- Self-hosted or fine-tuned models
- Model/API token costs
- Production build (quoted after)
Fixed prices, agreed before work starts.
One workflow or product area, up to 5 use cases scored, top pick detailed. Fee credited toward a build of $1,000 or more approved within 30 days.
Up to 3 areas, up to 10 use cases, tests on real examples for the top 2
Standard plus a working prototype of one use case on your sample data. Billed in 2 monthly milestones of $880 ($660 founding).
Build your scope
- Timeline
- ~1 week
- Payment
- Paid in full before work starts
- Earliest start
- Oct 5
The fee is credited in full toward a follow-on project of $1,000+ approved within 30 days.
Not a commitment. I confirm the exact price in a written proposal.Founding price: 3 of 7 spots left, until Mar 31, 2027 — one-off work, in return for a testimonial. How it works
Prices are in USD, paid by bank transfer against an invoice. Work under $1,000 is paid upfront; larger work is 50% to start and 50% before handover, or monthly milestones. The full fee of this service is credited toward a follow-on project of $1,000 or more that you approve within 30 days.How working together works
Variations
- Operations workflow
- For catalogs, documents and email, the focus is on time saved per item and the review step.
- SaaS feature
- The focus is user value, per-tenant cost and failure handling in the UI.
- Existing pilot
- I review the prompts, validation and failure cases first.
How it runs.
- 01
Email the template
Email the template below, and I reply within 1 US business day.
- 02
Prepay and share
You pay a prepaid invoice by bank transfer, then send examples and access.
- 03
Score and test
I score, test and write up the findings.
- 04
Delivery and walkthrough
I deliver the report with one walkthrough call or recorded video.
- 05
Optional build
If you want, I build the top pick as an AI Features project.
Where I have done this before.
AI Listing Pipeline for a Luxury-Watch Auction House
An LLM writes validated auction listings under rules that ban unsupported claims, with 1,027 listings since May 2026.
Read the case study →Multimodal AI · Auction catalogingAI Cataloging System for Fine-Jewelry Auctions
Two Gemini vision stages that are forbidden from guessing stone counts, carats or metal purity. Those facts come from supplier data only.
Read the case study →Questions buyers ask.
Will AI replace staff?
I don’t promise that. The usual realistic outcome is that staff move from writing to reviewing.
How accurate will it be?
That’s measured on your examples during the test, and you validate the results.
Which models do you use?
Hosted APIs (for example Gemini on Vertex AI, or providers through OpenRouter), chosen per task and cost.
Is our data sent to the model provider?
Only the samples you approve, and the report shows each provider’s data terms.
What does it cost to run?
The report estimates monthly token costs at your volume.
Ask about AI Opportunity Assessment.
Tell me what you have and where it hurts. I reply within one US business day with either a fixed quote or the entry step that makes a fixed quote possible.
Helpful to include
- Business or product
- The workflow or feature in question
- Rough volume (items per week)
- Current tools
- Sample data available (yes/no)
- Budget range
- Deadline
- Who approves
More in Discovery & Technical Planning
Discovery Sprint
A Discovery Sprint is a fixed-price, written plan for a software idea or a big new feature, made before you commit a build budget. It covers feasibility, stack, build vs buy and an estimate range. It is for founders and small teams. Prices start from $390 ($300 founding), and Starter takes about a week.
Second Opinion on a Quote
A new service: a fixed-price, written review of a developer or agency proposal before you sign. It covers scope gaps, risky assumptions, whether the price and timeline are realistic, who owns the code and accounts, and the questions to ask. From $130 ($100 founding), within 3 working days.