We delivered an AI Chief Marketing Officer for the owner of a Charleston roofing company — one that runs on his own laptop on a schedule, checks where his company stands in the local market every morning, and adjusts his listings, reviews, ads and follow-ups before he has finished his coffee.

He started the company in 2011 with one crew, one dump trailer and a Ford F-250 that is still in the yard. Fifteen years later it is four crews, about 380 roofs a year, and a name that people in Summerville and West Ashley recognize on a truck. He is good at roofs. He has never been good at marketing, and for most of those fifteen years he did not have to be, because storms did the marketing for him.

That stopped being true around 2023. Three national consolidators moved into the Charleston market with call centers and ad budgets he could not match. His phone still rang after a hail event, but between storms it went quiet in a way it never had, and when he searched his own trade on his own phone, his company came up fourth, behind two of the consolidators and a firm in North Charleston that had been open for eighteen months. He had paid two marketing agencies over the years.

Both sent a monthly report he did not read and both were, in his words, "renting me a website." What he wanted was a person. Someone whose job was to look at the whole Charleston market every day and move his company toward the front of it, and who would tell him plainly what they had done. The AI Chief Marketing Officer runs as Claude Cowork on the laptop in his office.

It does not live in a cloud dashboard he has to log into. It lives on his machine, with his files, and it runs on a schedule. Every morning at 6:15 it goes to work. It checks how the company ranks for the two dozen searches that matter in this market, from Mount Pleasant to Johns Island to Goose Creek, and notes what moved.

It reads every new review, his and his competitors', and drafts a reply to each of his that he approves with one tap. It looks at what the consolidators changed on their listings and ad copy in the last 24 hours, because they change things constantly and most of it is a signal. It checks his Google Business Profile for the things that quietly break, like a wrong holiday hour or a photo that got flagged, and fixes them. It reviews the ad spend from the previous day, pauses what is not converting, and shifts the budget toward the neighborhoods where his crews are already working, because a truck on the street is the best ad he has and the AI CMO knows where the trucks are from the job calendar.

By 6:40 he has a note. Half a page. What changed in the market, what it did about it, what it wants permission to do, and one number: where his company ranked that morning across the whole set of searches, averaged. The first morning it was 4.

3. He watches that number the way he used to watch the radar. Twice a week it goes further. It writes and schedules the content that keeps a local company visible between storms, a short piece on what a homeowner south of Broad should know about a slate repair, or why an architectural shingle warranty on Daniel Island reads differently than it does inland.

It drafts a follow-up to every estimate that has gone quiet for more than five days, in his voice, and sends it once he approves. It keeps a list of the 61 neighborhoods in his service area and knows which ones he has not worked in ninety days. He did not expect the number to move quickly. It did.

The average rank went from 4. 3 to 2. 1 in eleven weeks. Estimate requests from search, not storms, went from about nine a month to 27.

His ad spend went down, not up, by about 30 percent, because most of what he had been paying for was being wasted on searches from outside the Charleston metro. And he stopped paying the agency. He still does not read marketing reports. He reads half a page at 6:40 in the morning, and then he goes and looks at roofs.

"It's the first marketing person I've ever had that actually works for me. It's on my laptop. It's mine."