Stop asking AI questions.
Start putting it to work.
Five ways to work together, for small businesses that have heard plenty about AI and gotten very little that was useful. Pick the one that matches where you are.
Work that took my predecessor a full week, I get done in a day. Some of it now takes five minutes.
I work for a statewide trade association. Small staff, and the work never runs out. Putting AI to work there changed how we operate, and I have been at it long enough to know what actually sticks and what gets abandoned in a month. Now I do the same for other small businesses.
Most of what gets sold as AI strategy is a demo of somebody else’s software. What actually helps is far less exciting: figuring out which repetitive jobs in your week are worth handing over, which ones absolutely are not, and what you are already paying for and never opened.
That is what the audit is for, and it is where nearly everyone should start.
Operations audit
I spend half a day watching how your week actually runs. Not a questionnaire. What gets typed twice, what sits waiting on one person, what’s held together with a spreadsheet and sticky notes. Then I tell you what AI is genuinely good at here, what isn’t worth paying for, and what you’re already paying for and not using.
- Written findings, ranked by payback
- Specific tools, with real costs
- What to skip, and why
- A 90-day sequence, not a wish list
Employee training
A working session with your staff, on your systems, using the jobs they actually do. Not a talk about the future of work. What to hand over, what to never paste in, and how to check its work before it goes out with your name on it. Built for people who are busy and a little skeptical, which is most people.
- 90 minutes to a half day, your call
- A one-page reference for each person
- Ready-made wording for the jobs they do
- Follow-up questions answered for 30 days
AI use policy
Once your people start using this on their own you’ve got a policy question, whether you’ve written one down or not. What’s allowed, what isn’t, who signs off on a new tool, what happens to customer information. The answer looks different for an insurance agency than it does for a landscaping company.
- A policy short enough to be read
- Written for your industry’s obligations
- An approved-tools list you can update
- A rollout you can do in one meeting
Getting found by AI
More and more, people ask an AI instead of searching Google, and it answers out of whatever it can read and trust. Keywords don’t decide that anymore. I look at what AI says about your business right now, then rebuild your content so you’re the thing it quotes. It’s what I did for our association’s consumer site.
- What AI says about you today
- Pages written to be found and cited
- A chatbot that answers on your site
- Tracking, so you know it worked
Custom builds
For the repetitive job nothing off the shelf handles, where your workaround right now is a spreadsheet and a stack of paper. I scope it with you, write the spec, and build it. Honestly, deciding how it should work is most of the effort, and we do that part together before anything gets built.
- A written spec before any cost
- A working tool, not a prototype
- Fixed price, agreed up front
- Handover so you’re not dependent on me
Findings, ranked by payback
This is the shape of what the audit hands you. Every row is something a person on staff does by hand every week, measured first and then judged. Figures below are illustrative.
| What I found | Hrs / wk | Verdict | First step |
|---|---|---|---|
| Quote follow-up letters retyped from scratch | 6.5 | Worth it | Drafted from your last 200 sent letters, with a person approving every send |
| Intake calls typed into the system afterward | 4.0 | Worth it | Transcription straight into the fields your staff already use |
| Answering the same twenty questions by email | 5.5 | Worth it | You already pay for this. Copilot is sitting in the corner of Outlook |
| Monthly commission reconciliation | 3.0 | Skip it | Too many exceptions. Automating this moves errors somewhere nobody will catch them |
Three of four are worth acting on. The fourth is the reason to have someone say so out loud before you buy anything.
How this usually starts
With the audit. $1,500, flat. If you book any of the other work within 90 days, that $1,500 comes off the price of it. It’s the cheapest way to find out whether the rest is worth doing, and it’s the only part that has to happen in your building. What comes after is whatever the audit turned up, quoted as a flat fee before we start. No hourly, no surprises.
Questions I get asked
What does an AI operations audit cost?
$1,500, flat. That covers the half day on site, the written findings, and the 90-day sequence at the end of them. If you book any of the other four services within 90 days of the audit, the $1,500 comes off that price.
How long does it take?
Half a day in your building, and written findings back to you inside a week. You don’t have to prepare anything or fill out a questionnaire beforehand. The point is to watch a normal day, not a staged one.
Where do you work?
I’m in Pierre and I drive. Rapid City, Sioux Falls, Aberdeen, Watertown, Brookings, Mitchell, Yankton, Huron, Spearfish, and the towns in between are all a normal day for me.
The audit is the only part that has to happen in person. Training, policy work, and builds are done remotely unless you’d rather I came out.
What kind of businesses is this for?
Small ones. Two people to about fifty. Insurance agencies, contractors, clinics, law offices, ag operations, municipal offices, trade associations. The common thread isn’t the industry. It’s a small staff doing the same handful of repetitive jobs every week and nobody with time to go figure out whether software could take some of it.
Do I have to buy new software?
Often not. One of the most common findings is that you already pay for it and never turned it on. If you have Microsoft 365, Copilot is sitting in the corner of Outlook right now. Where something new is genuinely worth buying I’ll say so, with the real monthly cost next to it.
What about our customer information?
That’s the whole reason the AI use policy exists. Some of your information can go into these tools and some of it absolutely cannot, and the line sits in a different place for an insurance agency than it does for a landscaping company. Getting that written down before your staff starts experimenting on their own is the cheap version of this problem.
What if AI isn’t right for my business?
Then I’ll tell you that, and you’ll have it in writing. The findings mark things skip it as often as they mark things worth doing, usually where automating something would move errors to a place nobody would catch them. Knowing what not to buy is most of what you’re paying for.
Do you speak to associations and member groups?
Yes, and it’s some of my favorite work. A session built for one industry, using that industry’s actual jobs, for a conference or a member meeting. I work at a statewide trade association myself, so I know how a program committee books and what a room of members will and won’t sit through. Get in touch with your dates.
Worth a conversation?
Email me whatever’s eating your week. If I can’t help you, I’ll tell you that too.
Get in touch