← FIELD GUIDEFIELD NOTE / 05
First workflow8 min read

Send Me a 7-Minute Walkthrough of Your Shop. I'll Tell You Where AI Pays.

For an AI assessment for local business, skip the org chart. Film a 7-minute walkthrough of one job and I'll score where AI pays.

The owner of four hardware stores is standing in the back of store two with his phone up. Not filming the new endcap. Filming the complaint binder, the manager's personal texts, and the inbox that is supposedly "corporate" and is actually him. A customer asked store three for a raincheck on a SKU store one still had on the shelf. The answers didn't match. Sunday night he will build the out-of-stock sheet by hand again.

I'm not going to ask him for an org chart, a tech stack spreadsheet, or a 40-slide vision. I need to watch how a job actually moves. Seven minutes is plenty. Front desk. Dispatch. Inbox. The path a new patient, a service call, a consult, a lead, or a store complaint takes from "it showed up" to "we closed it or we dropped it." Talk while you walk. Point at the screens. Point at the paper. Point at the group text you wish wasn't the system.

I will map three paid opportunities — scored on money, time, and the customer's experience — and put them in a short assessment you can use. If you want, I will also record a short video of your report so a spouse, partner, office manager, or outside advisor can watch it without sitting in on the call. That's the offer. It's paid. It's the $1,500 assessment: a 45–60 minute owner interview and seven-business-day analysis plus the walkthrough you already live. It is not a free look.

What I need you to film

Not a tour of the lobby. Not the new chairs. The job.

Dental: start when the phone rings or the recall list prints. Who picks up. What they type, and into how many places. Where the schedule disagrees with the card. What happens when insurance is a mess. What happens when a hygiene opening appears for Thursday. If the office manager is the plot, follow her, not the dentist.

HVAC: start at the missed call, the voicemail light, or the estimate that still isn't out. Dispatch board, whiteboard, group text, the tech in the driveway. How a diagnostic becomes a quote. Where the homeowner sits in the dark. Where you, the owner, still get pinged on Saturday.

Law: start at inquiry. Web form, voicemail, receptionist note. How it becomes a consult. What the follow-up looks like the night after. Where facts live. Where they die. The managing partner's version and the intake coordinator's version are often different. Film the person who actually does it.

Agencies: start at the call that should have become a proposal. Recording, notes, the blank deck, the Slack thread. How long "I'll send something tomorrow" actually takes. Where claims sneak in that you would never put on paper if you were awake. Multi-location retail: start at store two's complaint, or the Sunday night report, or the review that sat for four days. How store three hears a different answer than store one. The manager's phone. The inbox. The binder.

Seven minutes will not capture a year. It will capture the truth better than a form. People describe the SOP. Cameras catch the extra step. If you don't want to film the floor, a screen recording of the inbox and the PMS, CRM, or dispatch tool while you narrate is enough. Same length. Same job. Don't perform. Don't clean up the desktop for me. The mess is the point.

What I am looking for

Frequency. If it happened once, it's a story. If it happens every day, it's a candidate. Handoffs. Every extra person is a place the job stalls. Dental front desk to coordinator to dentist's cell. HVAC CSR to dispatch to tech to estimator to owner. Law receptionist to intake to attorney. Agency AE to PM to whoever writes. Retail cashier to manager to "corporate" inbox that is actually you.

Copy-paste. Two systems, same facts, typed twice. That's Efficiency sitting on the table. It is also Quality, because the second typing is where the error lands. Places a first draft would help a human. Recall texts. Estimates. Follow-ups. Proposal outlines. Review replies. Shift notes. The human still sends. I will not recommend a bot that emails patients, clients, or homeowners on its own. I will not recommend fabricating a Google review. Ever.

Places the process is longer than the work. Fifteen steps that should be eight. I will tell you to cut before I tell you to automate. A faster mess is still a mess. Owner-visible numbers. If I can't tell you what we'd look at in two weeks — filled chairs, quotes out, follow-ups sent, complaints closed — I'm not going to dress it up as an AI project.

Hypothetical (not a client, not a promise): a four-store operator films seven minutes of Sunday-night reporting. Three hours of copy-paste, plus customers driving to a store that doesn't have the SKU. If even two wasted trips a week were $40 baskets you already advertised, that's a napkin on Effectiveness. If the three hours come back to the owner, that's Efficiency — but only if he uses them. Hours are not dollars until the shop uses them. I will not invent a "retail transformation" number. Count your wasted trips.

What you get back: three opportunities, not a catalog

I don't need to list twenty tools. I need three that pass the filter.

Effectiveness: it makes money. Booked diagnostics from calls you already missed. Hygiene openings filled. Consults that don't go cold. Proposals out while the lead is warm. A complaint that doesn't become a lost customer and a one-star that sits unanswered. Efficiency: it gives hours back to a named person. The coordinator's morning. The estimator's evening. The paralegal's intake pile. The store lead's Sunday. Quality: the customer can feel it. Complete estimates. Intake they don't repeat. Replies that sound like your shop. Same answer at every location.

Each opportunity gets a plain card: what I saw in your walkthrough, what we'd change (process, off-the-shelf tool, or both), who would own it on Tuesday, what we'd measure, and a conservative sketch of the leak. Hypothetical math, assumptions listed. Not a case study. Not someone else's hours. Your shop, on a napkin, honest about what we don't know.

I will also tell you what not to do. The voice agent that books the wrong window. The custom build you don't need. The "AI receptionist that also does billing." The seats you already bought that aren't attached to a loop. That "no" list is part of what you pay for.

The written assessment is for you and for the walkthrough call. The optional video is for the rooms I will never be in. You run a practice with a spouse who handles the books. A firm with other partners. An HVAC company where the operations person wasn't on the Zoom. They will not read a PDF at 9 p.m. They will watch six or seven minutes if it opens with your shop's name and walks first, second, third. I will not end that video on a made-up annual miracle. I will end on the same conservative model that's on the page, labeled as a model.

How this is not a free audit

Do not send a cold video and ask me to "take a look." That's the unpaid audit I refuse. You buy the assessment. I send you what to capture (one job, seven minutes, narrate the ugly parts). You send the clip, plus a few facts I need in writing. We get on a working interview if the video isn't enough. I write the scored map. We walk it. Optional: the short video for the partner who wasn't there.

$1,500. Paid before I watch. If you only want the filter, not a map of your shop, the free 15-minute Mini Margin Check is the room for that. No filming required. If you want neither, you can keep the folder of vendor PDFs.

A good seven minutes sounds like this: "This is the recall list. She prints it because the software's 'overdue' view lies. Then she checks the yellow sticky, then the schedule, then she texts from her personal phone because patients answer that. Thursday's 2 p.m. has been open for nine days." Or: "This is the after-hours box. I listen in the truck. If I don't, nobody does. Half of these are 'my AC died' and we call back at 8:12 when they're already on someone else's site." Polished brand video is not gold. That is.

Use this as the voiceover checklist so you don't perform:

Walkthrough script (talk, don't clean up):
1. This is the job: [recall / missed call / intake / proposal / complaint]
2. It starts when: [phone / form / voicemail / list prints]
3. Who touches it, in order:
4. Where we type the same facts twice:
5. Where it dies (be honest):
6. What I, the owner, still get pinged for:
7. The number I can see in two weeks if this got better:
Do not film the lobby. Do not sell me. Point at the mess.

I will not post your clip. I will not use it as a testimonial. I will not train a public model on it. I will not let a helper send anything to your patients, clients, or customers from what I saw. Human review, always, on anything outbound. I will not turn seven minutes into a custom software quote before the map exists. Most first wins are a cleaned path plus a helper with your rules, used by one person.

FAQ

What do you need for an AI assessment for a local business? A paid 45–60 minute owner interview and seven-business-day analysis and a seven-minute walkthrough of one real job. Org charts and vision decks can stay in the drawer. I need to see how work actually moves.

Can I email a video and ask you to take a look for free? No. A cold clip with "take a look" is the unpaid audit I refuse. Buy the assessment. Then I tell you exactly what to film.

What if I don't want cameras on the floor? Screen-record the inbox and the PMS, CRM, or dispatch tool while you narrate. Same seven minutes. Same job. The mess on the desktop is the point.

If you're the dentist, HVAC owner, lawyer, agency lead, or multi-location operator who is tired of describing the shop in adjectives, film the job and buy the assessment. $1,500 for a 45–60 minute owner interview and seven-business-day analysis. I'll tell you where AI pays — and where it doesn't — in three items you can act on. I don't need your org chart. I need the seven minutes you already live.

$1,500 · 7 BUSINESS DAYS

Turn the ideas into a ranked decision.

The Margin Scan compares three-to-seven opportunities on impact, cost, effort, and risk, then names the first move—or recommends leaving it alone.

Apply for a Margin Scan