The three practices are a sequence, not a menu. You are on step 1 of 3.
AI Business Transformation
Find the work AI should be doing. Build it. Keep it running.
Most small businesses are being sold AI they do not need, by people who have never run a business like theirs.
The problem, in the operator's words
You are being pitched tools that solve problems you do not have. The demo looks impressive, but nobody can tell you what it will save you in hours, dollars, or stress.
I have run small businesses. I know that a new dashboard is not a strategy, and that the best AI project is the one your staff actually use.
My job is to find the one or two places where AI makes your business faster, safer, or more profitable - and to ignore the rest.
What I actually do
Three steps. No jargon. Every build starts with a free audit.
1. Audit
I start with a free Signal Audit. In a few minutes you see where AI is risky in your business, where it can pay off, and what to do first - with a Signal Score.
2. Build
Then I build the best safe win: a customer "front desk" that answers questions and takes bookings, or a back-office automation that kills a repetitive task. On tools you own.
3. Watch
Then I keep it safe and current. When tools and risks change, I update your setup and tell you what matters.
My method: detect · decide · deploy · defend
What gets built
Four concrete automations I build for small businesses. Every claim carries a source or a status label.
Front-desk automation
An AI front desk that answers common questions, books appointments, and routes urgent calls.
Replaces: [placeholder - needs source] hours a week of repeated phone and message handling
Back-office automation
One repetitive workflow configured end to end: invoicing, scheduling, reporting, or lead follow-up.
Replaces: [placeholder - needs source] admin hours per week
Document and intake handling
Forms, contracts, and intake documents read, sorted, and filed automatically - not by hand.
Replaces: [placeholder - needs source] hours of manual data entry monthly
Internal knowledge search
A search layer over your SOPs, files, and past answers so staff stop asking the same questions.
Replaces: [placeholder - needs source] hours a week spent tracking down information
The Loonie Test
If it does not show real ROI on the P&L within 90 days, it was noise.
I apply this to every project I recommend. That means I define a clear success metric before I build - hours saved, errors reduced, revenue protected, or cost avoided - and I check it after 90 days. If the numbers do not hold, I fix it or I kill it.
What you keep
Everything runs on tools you already own or on standard business accounts set up in your name.
You own the accounts. You own the data. You own the solution.
If you leave, you keep what was built. No lock-in. No ransom. No surprise migration fees.
Start with a free Signal Audit.
Eight minutes. See your risks and opportunities, and get a clear next step.
Take the free auditThe product for this step
This page describes the work. Here is the product that does it, with the price, the scope, and what is not included.
Signal Plan
Two weeks, I do it
2 of 4 August engagements left.
Everything AI should be doing in your business, ranked by payback - and the first one built and running before I leave.
In scope
- One workflow end to end, front desk or back office
- A runbook
- A 12-month roadmap
- A re-scored Signal Score
Not in scope
- New software purchases without your sign-off
- Integrations to systems you do not already own
Proof
Hours, dollars or errors on that workflow - baselined Day 1, measured Day 30.
Where this price comes from
About two of my days. An agency would quote six weeks and $20,000. It does not take that any more, so I do not charge that.
Next: step 2 - Make it repeatable . Once the work is found and built, it only pays off if anyone can run it twice. Step 2 makes it repeatable.