Bring us the work that is hard to get right.

The records that never quite agree. The task that eats up an afternoon. The business question your reports don’t answer. You don’t need an AI project in mind.

Book 15 minutes with Billy
We’re accepting new clients.
or call (817) 265-5000 Or ask Billy a question
Billy at a workbench directing several AI models and task agents, with research, document organization, record matching and software work feeding into human review.
Billy directs the AI models and task agents. People review the results.

Different businesses.
Different work to improve.

A few examples of where we’ve put AI to work.

Law

Years of documents, put in order.

Emails, contracts and other records became a draft chronology with facts linked to the original document and page.

A first draft in about 35 minutes instead of about six hours, ready for attorney review.

Med spa

Missed calls became booked visits.

We connected an AI receptionist to the spa’s scheduling and service information. It answered questions and booked eligible appointments, passing clinical questions to staff.

About 60 completed visits a month came from calls previously missed.

Construction

Extra work that missed the invoice.

Project emails and field reports were matched against change orders and job-cost records.

The contractor reviewed the findings, got the extra work approved and billed, and collected about $148,000 that had gone unbilled.

Property management

Hidden leaks became repair jobs.

Meter readings and sensors flagged unusual water use. AI matched alerts with maintenance history and prepared work orders.

Seven leaks were confirmed during a six-month pilot.

Nonprofits

More grant applications. Same small team.

AI read funding requirements, drafted sections from verified organizational information and assembled document checklists. Staff reviewed every submission.

A six-person team completed 18 applications in a year instead of seven.

Churches

The fix didn’t depend on one volunteer.

Alongside standardized systems and monitoring, an AI assistant used approved documentation to guide volunteers through common technical problems.

Routine fixes fell from about 25 minutes to eight during a six-month pilot.

Closing two offices looked like the answer.

Three offices. Two years of decline. The owner was considering closing two locations. We looked at what was actually going wrong.

We pulled data from their own systems and used AI to connect information that would have taken months to piece together by hand, if it could have been done at all. The answers were hidden in those records. Bringing them together revealed where the business was failing and what needed to change.

  1. The question

    Would closing the offices fix the business?

    The business was no longer generating enough to cover what it paid its owners. The reports made closure look like the obvious way to cut costs.

    • Clients were drifting toward cheaper, easier-to-access competitors.
    • Paid staff capacity was sitting unused.
    • The owner needed to know which costs would actually disappear if the offices closed.

    The real questionWhat would closing save, and what would it lose? We examined the underlying financial and operating records, then tested the analysis against how the business actually worked.

  2. What we found

    The failures went well beyond the accounts.

    An accounting convention distorted the location comparison. Closing offices would save less than the reports suggested. Meanwhile, opportunities were being lost before the team even spoke with a prospective client.

    • Contact-form messages were not reaching a monitored inbox.
    • Contact details conflicted, and booking and location pages were missing or hard to find.
    • Returning-client outreach was underway, but its results were not being measured.

    What needed fixingLost inquiries. Confusing contact details. Unused capacity. The business had useful work to recover and room to deliver it with people it was already paying.

  3. What changed

    Ninety days of changes. A year of follow-through.

    We helped repair the contact and booking problems, clarify the offer, measure returning-client outreach, and test better appointment access and referral relationships. Then we kept measuring and refining the work.

    • The first ninety days established the improvements and how to measure them.
    • The following nine months continued the work and adjusted what was not working.
    • The result came from better use of the existing business, without opening another office.

    Twelve months laterThe offices stayed open. Revenue had grown, and the business had money left after paying its owners, from its existing locations.

How we work on this

We take no vendor commissions. We start with the tools you already have, check what your licenses include, and recommend what fits. We show you what is possible, talk through what matters to your business, and begin with the improvement that would make the biggest difference.

For our ongoing clients, AI advice and implementation are generally part of the relationship. We usually build the improvement, help your people use it, and keep looking for what could work better.

How we charge
Billy Dunn, illustrated
Billy DunnFounder. AI problem-solver.

Studying AI and experimenting independently since the 1980s.

Good AI starts with good judgment.

First, the models challenge each other.

We put different AI models to work checking one another’s reasoning, evidence and conclusions. They look for errors, unsupported claims and invented facts. We work through the disagreements against the underlying records until we have high confidence in the result. Agreement between models is a checkpoint; the evidence still has to hold up.

Then, they work together.

We have the models collaborate to refine the answer, connect useful findings and develop a more complete result. We direct the process, bring your knowledge of the business into the review, and turn that work into something your business can use.

Protection is designed into the work.

We build in checks and limits that protect your systems, data, and customers or patients. They control what the AI can see, what it can change, and when it must stop and ask.

The relationship behind the improvements

I was introduced to BDunn & Company when looking for a new IT option; they were referred to as "computer wizards". They've lived up to that description.

Cammie S., Google review

Every review on Google: five stars.

More reviews

What would you like to stop doing the hard way?

Billy reads these himself. You’re welcome to ask while you’re still exploring, including before involving the owner.

Add your contact details

Either one is fine.

Best way to reach you

During office hours, we usually reply within an hour.

No mailing lists. No sales pitch. Just a reply from us.

Call Book 15 minutes with Billy