AI Opportunity

Figure out what AI is actually worth doing in your business.

In 10 business days, I’ll look across your business for places AI and related technology could create meaningful value, quantify the strongest opportunities, and tell you what to pursue, what to postpone, and what to ignore.

$5,000 fixed fee10 business days

Schedule Our First Chat

Free 30-minute conversation. Scope agreed before paid work begins.

What you get

The working register: Ranked opportunities, assumptions, economics, dependencies, and recommendations in an editable file.
The decision brief: A concise account of what matters most, what to leave alone, and the first experiments worth running.
The walkthrough: A conversation with me to challenge the findings and decide who should do what next.

The details

How the engagement works

Understand the business. We start with your goals, economics, and constraints. We agree the areas to examine and the people and information I’ll need.
Find the opportunities. I talk with the people doing the work and inspect workflows, systems, and examples. We look for value across the agreed scope.
Quantify and pressure-test. I estimate what the strongest opportunities are worth, what they require, and what could go wrong. Assumptions and missing evidence stay visible.
Make the calls. We review the opportunity map together: what to test first, what needs more evidence, and what to leave alone. You keep the working files.

The 10 business days start at kickoff, once the agreed information and interview access are available. We agree the areas to review within one business or business unit before starting; if access is delayed or scope changes, we agree a revised schedule. Production builds and pilots are separate work.

No software purchase, implementation contract, or continuing engagement is required.

When this is a good fit

You run an established business with repeat work, expensive bottlenecks, or room to serve customers better.
You want to know where AI is worth the investment before committing to tools or a build.
You can involve the people doing the work and share examples of how it happens.
Someone has the authority to act on the findings, including a recommendation to stop.

When something else would help more

You already have a defined specification and need someone to implement it.
You want a general AI training session or a list of tools to try.
You need someone to validate a purchase you have already decided to make.
Nobody has time to show the work or responsibility for what happens afterward.

See an example opportunity map

Fictional business. Assumed figures. This shows the format of the assessment, not client findings or promised results.

Proposal developmentTest now

Senior staff reconstruct proposals from prior work. Test an AI-assisted first draft using approved material, with a person responsible for every final proposal.

Current annual labor cost: ~$87K20 hrs/week × $87/hr × 50 weeks
Capacity hypothesis: 8 hrs/week~$34.8K/year of staff time, if the test supports it

Value mechanism: Senior capacity and faster proposal turnaround. Freed time only creates value if the business can use it.
AI fit / difficulty: High / medium, subject to a test of draft quality and review effort.
Dependencies: Clean past proposals, approved qualifications and language, and a named reviewer.
Costs to establish: Tool licenses, content cleanup, setup, human review, and ongoing maintenance. Obtain a quote before committing to production.
Main risk: Invented qualifications or pricing. Keep pricing outside the draft generator and check every factual claim.
Build or buy: Test an existing tool with approved source material before considering custom software.
First experiment: 20 historical proposals, then a 30-day pilot. Compare total drafting and review time, factual errors, and approval rates against the current process.
Decision gate: Proceed only if total effort falls without worse quality. Otherwise revise the approach or stop.

All figures above are illustrative assumptions. Labor cost is not recoverable savings, and capacity value is before implementation and running costs. The pilot follows the assessment; it is not included in it.

Answers from internal knowledgeInvestigate

People spend time finding answers, but the source documents disagree. A faster answer is not useful if it is wrong.

Before a pilot: Name the content owner, reconcile conflicting guidance, and measure how often the questions recur.
Value / costs: Unquantified until search time, question volume, and content maintenance effort are measured.
Recommendation: Investigate the information problem first. Test AI retrieval only after the source material can support reliable answers.

An AI agent for approval routingDon’t bother

Approval rules are fixed and already supported by the existing system. The delay comes from missing ownership.

Better next move: Assign an owner and configure the existing approval rules.
AI advantage: None established. An agent adds ongoing cost and another failure point.
Recommendation: Fix the handoff. Measure the resulting turnaround before spending on another tool.

Related work: the ERP decision

One of the most valuable technology recommendations I’ve made was to stop implementing the technology.

A PE-backed distributor had purchased an ERP that could not represent how the business actually worked. I recommended stopping. The CEO canceled the implementation, and I designed an alternative around systems the company already had.

The ERP license had already been contracted, and about $350K had been paid when implementation stopped. These are projected costs and a design estimate, not a claim of realized savings.

Read the ERP case

Worth the fee, or your money back

Complete the assessment and review the findings with me. If you don’t believe the decisions and recommendations are worth the $5,000 fee, tell me within seven calendar days of the review. I’ll refund the full fee. You keep the findings. Give me the access we agree on and join the findings review. You do not have to implement a recommendation or buy further work to use this guarantee.

What does my team need to do?

An owner or executive needs to sponsor the work, make introductions, and join the kickoff and findings review. The people closest to the work take part in focused conversations and show real examples. We agree those conversations and access before starting.

Do we need clean data or an AI strategy already?

No. Messy information and unclear priorities are often why this is useful. I do need access to enough real work to distinguish an opportunity from a guess. If a prerequisite is missing, the recommendation will say what needs to change first.

Will you build or implement the recommendations?

This engagement ends with the assessment, priorities, and first experiments. Implementation, software, and ongoing support are separate. You can take the findings to your own team or another provider; there is no obligation to hire me for anything else.

What if AI is not worth doing?

That is a valid result. The assessment will explain why and identify the better next move, which may be a simpler process, clearer ownership, a conventional integration, or no project. I do not receive vendor commissions or referral fees.

Are the estimates guaranteed savings?

No. I distinguish measured costs from assumptions, freed capacity from cash savings, and an attractive hypothesis from a tested result. The first experiment is designed to find out whether the value survives contact with the real work.

How do you handle sensitive information?

We agree access and handling before work begins, including what can be shared, what should be redacted, and which tools are approved. Do not send confidential business records when booking the introductory call.