Scale with AI / Episode 43 / November 15, 2026
Scale with AI — November 15, 2026
Marketing AI used to mean help writing a post. The new pattern is a working loop: the system studies your real sales data, drafts the ads and messages, a person you name approves them, the campaign runs, and you get a plain daily report on what sold. A working paper published this month found the largest gains went to the smallest firms. Plus the November 25 opt-out deadline — ten days out — and the three moves to make this week.

Quick Take
- The Shift: marketing AI is moving from writing posts to running a loop — study real sales data, draft the ads and messages, a named owner approves, the campaign runs, and a plain daily report shows what sold. Judged on revenue, not content volume.
- The evidence: a working paper this month on small online sellers using AI built into their daily software found adopters an estimated 51.6 percent ahead on monthly sales and 47 percent on orders versus matched non-adopters — with the largest gains at the smallest firms. One platform, one study: an estimate, not a promise.
- The Law: November 25 is ten days out. An informational opt-out can cover just that category; a marketing opt-out ends all marketing; your designated opt-out method must be clearly stated and actually work; honor within ten business days at the outside — same day, logged, with proof is the smart standard.
- Your Move: fix the funnel before buying traffic (buy your own product on your phone), automate exactly one marketing loop with the approval line written first, and run the opt-out test this week.
The Shift
Marketing AI used to mean help writing a post or an ad. That era is ending. The new pattern is a working loop. The system studies your real sales and customer data, drafts the ads and messages, a person you name approves them, the campaign runs, and you get a plain daily report on what sold, not just what got clicked. The work is judged on finished, approved campaigns and revenue, never on how much content it produced. Here is why that matters for a small shop. A working paper published this month studied small online sellers using AI built directly into the software they already ran their business on. Sellers who adopted it saw monthly sales an estimated 51.6 percent higher than similar sellers who did not, with orders up about 47 percent, and the largest gains went to the smallest firms. Honest limits: that is one platform, one group of sellers, and a research estimate, not a promise. But the pattern matches what we see. AI built into your daily work beats another subscription you have to remember to open.
The Fine Print
On November 25, ten days from today, new Federal Communications Commission rules take effect for how customers opt out of your automated calls and texts. The plain version: if a customer opts out after a purely informational message, like an appointment reminder, that opt-out can apply just to that kind of message. But if they opt out after any marketing message, that ends all of your marketing calls and texts to that person. Every business must clearly state its designated opt-out method, and that method has to actually work. You must honor an opt-out within ten business days at the legal maximum. Our advice is simpler: honor it the same day, log which category it was, and keep the proof. And remember the rule underneath all of this: since February 2024, an AI-generated voice counts as an artificial voice under federal law, so AI marketing calls carry the same consent duties as any robocall. A business that cannot prove consent and cannot prove it honored a stop is carrying a risk that dwarfs whatever the marketing earns. Test your own opt-out this week: text stop to your own business number, tell one of your own team members to stop your messages, and confirm both actually stopped and were logged. This is general information, not legal advice.
Who’s Doing It
Consider the pattern from a seafood shipper selling online across the country. An AI system analyzed its store and advertising data, wrote the ad ideas and creative briefs, produced the ads, and sent every one to the owner in a chat message for approval before a dollar was spent. In the first week, the owner approved about 20 percent of what the system made. By the end of the test, he was approving about 90 percent with no changes, because the system had learned his taste from his corrections. The reported test result was three and a half times the revenue of the company's usual ads, on about twice the spending — roughly 63 percent more revenue per advertising dollar. Honest limits again: one company, the figures are the platform's own report, and the test budget was provided for the trial. Do not expect those numbers. Do copy the structure: machine drafts, owner approves, nothing spends unapproved, and the approval rate itself tells you whether the system is learning your business. Compare that with a survey of 1,400 online sellers released this week: 59 percent use AI for marketing and social media, but only 15 percent use it for sales insight, 13 percent for customer service, and 4 percent for operations. Most businesses are using AI to make noise. The few using it to run the loop — from data to draft to approval to sale to report — are the ones pulling away.
Your Move
Three moves, in order. First, fix the funnel before you buy traffic. We saw a case this month with over 111,000 ad views and zero sales, because the page and the checkout failed on phones. AI that sends more people to a broken page just loses money faster. Open your own site on your phone and buy your own product before you spend anything on reach. Second, pick exactly one marketing loop to automate: product photos and drafts for one product line, or follow-up messages for one kind of lead. Write the approval line before you start — the system may draft anything, a named person approves anything that spends money or reaches a customer, and nothing with a price promise or a contract ever leaves unapproved. Third, run the opt-out test described above, because November 25 does not move. If you want help building that loop properly, with the approvals, the consent records, and the daily reporting set up for your business, that is exactly what we build at The Agency. Start with a free AI audit at theagencyllc.net, and we will show you which loop your business should automate first. The businesses winning with AI are not the ones with the most tools. They are the ones with one clean loop and a clear approval line.
Key Takeaways
- The loop is the product: data to draft to approval to sale to report. Businesses using AI only to make content are making noise; the loop users are pulling away.
- The approval rate is the learning metric: when an owner goes from approving 20 percent of drafts to 90 percent unchanged, the system has learned the business — that curve matters more than any single campaign.
- Nothing spends unapproved: the system may draft anything, but a named person approves anything that spends money or reaches a customer, and price promises or contracts never leave unapproved.
- November 25 opt-out rules: informational opt-outs can be category-limited; one marketing opt-out ends all marketing; the designated method must work; keep the proof.
- Fix the funnel first: 111,000 ad views produced zero sales through a phone-broken page and checkout. AI sending more people to a broken page just loses money faster.
Full transcript
Alex: Welcome to Scale with A I for Sunday, November fifteenth, twenty twenty-six. Today we are talking about marketing, because the newest evidence says something surprising. The businesses gaining the most from A I in their marketing are not the biggest ones. They are the smallest ones, and the reason is not the tool they bought. It is the loop they built. Stay with us, because we will also cover the opt out deadline that is now ten days away, and it applies to every marketing text and call you send.
Jordan: Let us start with The Shift. What changed in marketing?
Alex: Marketing A I used to mean help writing a post or an ad. That era is ending. The new pattern is a working loop. The system studies your real sales and customer data, drafts the ads and messages, a person you name approves them, the campaign runs, and you get a plain daily report on what sold, not just what got clicked. The work is judged on finished, approved campaigns and revenue, never on how much content it produced. Here is why that matters for a small shop. A working paper published this month studied small online sellers using A I built directly into the software they already ran their business on. Sellers who adopted it saw monthly sales an estimated fifty-one point six percent higher than similar sellers who did not, with orders up about forty-seven percent, and the largest gains went to the smallest firms. Now, honest limits. That is one platform, one group of sellers, and a research estimate, not a promise. But the pattern matches what we see. A I built into your daily work beats another subscription you have to remember to open.
Jordan: That brings us to The Law, and the clock is real on this one.
Alex: It is. On November twenty-fifth, ten days from today, new Federal Communications Commission rules take effect for how customers opt out of your automated calls and texts. Here is the plain version. If a customer opts out after a purely informational message, like an appointment reminder, that opt out can apply just to that kind of message. But if they opt out after any marketing message, that ends all of your marketing calls and texts to that person. Every business must clearly state its designated opt out method, and that method has to actually work. You must honor an opt out within ten business days at the legal maximum. Our advice is simpler. Honor it the same day, log which category it was, and keep the proof. And remember the rule underneath all of this. Since February twenty twenty-four, an A I generated voice counts as an artificial voice under federal law, so A I marketing calls carry the same consent duties as any robocall. None of this is legal advice, and a business that markets heavily should have its setup reviewed. But a business that cannot prove consent and cannot prove it honored a stop is carrying a risk that dwarfs whatever the marketing earns. Test your own opt out this week. Text stop to your own business number, and tell one of your own team members to stop your messages, and confirm both actually stopped and were logged.
Jordan: Now Who's Doing It. Who is getting real results from this marketing loop?
Alex: Consider the pattern from a seafood shipper selling online across the country. An A I system analyzed its store and advertising data, wrote the ad ideas and creative briefs, produced the ads, and sent every one to the owner in a chat message for approval before a dollar was spent. In the first week, the owner approved about twenty percent of what the system made. By the end of the test, he was approving about ninety percent with no changes, because the system had learned his taste from his corrections. The reported test result was three and a half times the revenue of the company's usual ads, on about twice the spending, which works out to roughly sixty-three percent more revenue per advertising dollar. Honest limits again. That is one company, the figures are the platform's own report, and the test budget was provided for the trial. Do not expect those numbers. Do copy the structure. Machine drafts, owner approves, nothing spends unapproved, and the approval rate itself tells you whether the system is learning your business. Compare that with a survey of fourteen hundred online sellers released this week. Fifty-nine percent use A I for marketing and social media, but only fifteen percent use it for sales insight, thirteen percent for customer service, and four percent for operations. Most businesses are using A I to make noise. The few using it to run the loop, from data to draft to approval to sale to report, are the ones pulling away.
Jordan: So what is Your Move for a business owner listening today?
Alex: Three moves, in order. First, fix the funnel before you buy traffic. We saw a case this month with over one hundred eleven thousand ad views and zero sales, because the page and the checkout failed on phones. A I that sends more people to a broken page just loses money faster. Open your own site on your phone and buy your own product before you spend anything on reach. Second, pick exactly one marketing loop to automate. Product photos and drafts for one product line, or follow up messages for one kind of lead. Write the approval line before you start. The system may draft anything, a named person approves anything that spends money or reaches a customer, and nothing with a price promise or a contract ever leaves unapproved. Third, run the opt out test we described, because November twenty-fifth does not move. If you want help building that loop properly, with the approvals, the consent records, and the daily reporting set up for your business, that is exactly what we build at The Agency. Start with a free A I audit at theagencyllc dot net, and we will show you which loop your business should automate first. The technology is ready. The businesses winning with it are not the ones with the most tools. They are the ones with one clean loop and a clear approval line. We will see you tomorrow.
Sources
Your next conversation