Use AI in sales and marketing for the work around the conversation, not the conversation itself. In a small team that means four jobs: preparing for a sales call, turning call notes into a same-day follow-up, turning one real customer story into a week of marketing, and drafting proposals from the ones you have already won. Let AI draft as much as you like, but let nothing go out that you could not back with a receipt: a customer's own words, a real result, a written policy. That single line protects the thing a small firm actually sells, which is trust.
It matters now because marketing is where most small businesses already use AI, and it is where the gap between what owners assume and what customers feel is widest.
Where are small businesses already using AI in sales and marketing?
Marketing is the first place most small businesses put AI to work.
Intuit's 2026 AI Impact Report, built from more than 34,000 survey responses from small and midsize business owners in the US, Canada, the UK and Australia, says it in one line: "Adoption is highest in marketing, admin, and customer service. It's lowest where human judgment matters most." The same report found that 43% of US businesses say AI has increased their revenue, and only 2% say it has gone the other way (Intuit QuickBooks, May 2026).
So the live question for most owners is no longer whether to use AI in marketing. Someone on your team is already asking it for Instagram captions on a Friday afternoon. The question is which jobs it should do, and what it must never be allowed to say.
Do customers mind AI-written marketing?
Some do, and the people who make the marketing consistently underestimate how many.
In January 2026 the Interactive Advertising Bureau (IAB), the US advertising industry body, published "The AI Ad Gap Widens". It compared 104 US advertising executives with 505 Gen Z and Millennial consumers. 82% of the executives believed those consumers felt very or somewhat positive about AI-generated ads; 45% of the consumers actually did. The gap had grown from 32 points in 2024 to 37 points in 2026. Asked whether AI-made ads make a brand seem forward-thinking, 46% of executives said yes, against 22% of consumers (IAB, January 2026).
The same study holds a finding that points the other way. 73% of the consumers said that knowing an ad was made with AI would either increase their likelihood to buy or make no difference. The sample is younger US consumers only, so treat it as a signal, not a verdict on your customers.
My reading is simple. Most customers are not objecting to the tool. They object to what the tool produces when nobody is watching: the glossy, could-be-anyone copy that tells them no one at the business cared enough to write to them. For a big brand that costs a little attention. For a small firm it costs the one advantage it has. You do not win work on reach. You win it because the customer believes a real person at your business understands what they need, and generic copy spends that belief.
The Receipt Rule
The Receipt Rule is the line I suggest a small business holds when AI helps write its marketing and sales messages: AI may draft as much as you like, but it may not say anything you could not back with a receipt. Every claim, number, customer quote and promise in AI-drafted copy must trace to something the business can show: a customer's own words, a real result, a written policy or a product sheet. If there is no receipt, the sentence goes, however good it sounds.
Receipts come in four kinds, and between them they cover almost everything a small business says about itself:
- A customer's own words. A review they actually left, an email they actually sent, a sentence they have agreed you can quote.
- A real result. A job you finished, a figure from your own records, a before-and-after photo you can show.
- A written policy. Your guarantee, delivery times or returns terms, exactly as you have written them down.
- A product or service sheet. What the thing is, what it includes and what it does not.
Everything else AI produces is scaffolding: structure, phrasing, headlines, the fifth version of a subject line. Scaffolding is where AI saves the hours. Receipts are where the business keeps its word.
The rule is more than good manners; it runs close to the law on both sides of the Atlantic. In the US, the Federal Trade Commission's final rule on consumer reviews and testimonials bans reviews and testimonials "that misrepresent that they are by someone who does not exist, such as AI-generated fake reviews", and allows the agency to seek civil penalties against knowing violators (FTC, August 2024). In the UK, the Competition and Markets Authority's guidance under the Digital Markets, Competition and Consumers Act 2024 defines a fake review as "a consumer review that purports to be, but is not, based on a person's genuine experience", and lists reviews "generated by software applications (such as bots)" among the banned practices (CMA, April 2025). An AI-invented testimonial has no receipt by definition. I am not a lawyer, so check any claim you are unsure of with someone who is; the rule simply keeps you well back from the edge.
The Receipt Rule sits beside the Promise Rule I suggest for customer service, where AI may tell a customer anything the business has already written down, but it may never promise anything a person has not already decided. The Promise Rule governs replies. The Receipt Rule governs everything you send first.
Which sales and marketing jobs should AI do?
Give AI the jobs around the conversation, and keep people in the conversation itself.
These are the four I would start a small team on. Each one happens every week, takes real hours and already gets read by a person before it leaves the building, which are the three tests for a first job in the One Job Rule.
1. Preparing for a sales call
Before a first meeting, AI can read a prospect's website and recent public news and draft a one-page brief: what they do, what seems to be changing for them, and five questions worth asking. The reading still happens, but it starts from a page instead of a blank screen. The salesperson still decides which questions matter. Only public information goes in.
2. Turning call notes into a same-day follow-up
This is the job most small sales teams do worst, because it lands at the end of a long day. Dictate or paste rough notes, and have AI turn them into an email that restates what the customer said they need, what you agreed and what happens next. The receipt is the call itself, so every line must match what was actually said. The person who made the call reads it, corrects it and sends it. Customer notes go only into an approved business account, never a personal one, for the reasons set out in what to do about shadow AI.
3. Turning one real story into a week of marketing
Take one finished job or one happy customer and have AI turn it into a short case study, three social posts, a newsletter paragraph and an email to past customers. This is the best trade in small-business marketing: the story is the receipt, and AI multiplies it. What AI must not do is invent the story. Start from something real every time.
4. Drafting proposals from ones you have won
Give AI two or three past proposals that won work, plus the notes from the new inquiry, and ask for a first draft in the same structure. The person who owns the relationship then rewrites the parts that matter: the understanding of the customer's problem, the approach, and anything that commits the business. Every claim about what you have done before needs a real past job behind it.
What stays with people
The relationship. The judgment about which customers to pursue. Any claim about results. Anything sent in a named person's voice that they have not read. And every testimonial: AI can help you write a polite request asking a customer for a review, but it must never write the review.
What does this look like in a real small business?
Take a hypothetical 14-person kitchen and bathroom fitting company. Two people carry the sales and marketing: the owner, who does every home visit, and an office manager who handles quotes, follow-ups and the social media accounts.
Before, the office manager spends Friday afternoon staring at a blank post, and the owner writes follow-up emails on Sunday night, three days after the visit, when the customer has already had two other quotes.
Under the Receipt Rule the week changes in two places. After each home visit, the owner dictates two minutes of notes in the van. AI turns them into a follow-up email, the office manager checks it against the notes, and it goes out in the owner's name that afternoon. And every finished kitchen becomes the week's marketing. The fitter takes before-and-after photos, the customer is asked for one sentence about the job and whether it can be quoted, and AI drafts a short case study and three posts from those materials. Nothing is published that does not trace to a finished job, a real photo or the customer's own sentence.
What they stop doing is just as clear. Nobody asks AI to "write five great reviews for the website". Nobody lets it call the firm "the region's most trusted fitter". There is no receipt for either.
What I Tell Business Owners
I tell owners that the danger in AI marketing is not bad copy. It is plausible copy nobody checked.
"The hard part of AI at work isn't the technology. It's helping people use it well, safely and every day." That is the line on my business page, and marketing is where it bites first. Getting AI to produce words is easy. Getting a team to produce words it can stand behind is a habit, and habits have to be taught.
Marketing is also where many businesses first see AI earn its keep. In April 2023 I spoke about AI in marketing at a business event in Penrith. A year later, the marketing team at theidol.com was using AI for content writing, and the managing director asked me back to show teams across the business what was possible. That order is common: once one team has real work to show, the rest of the business starts asking how.
The instruction I hear most often is "make it sound more professional". It is the wrong one, because it strips out the specific details that make a small business worth choosing: the street name, the fitter's first name, the exact thing the customer was worried about. Try this standing instruction instead: "Keep every detail I have given you. Do not add any claim, number, quote or promise I have not given you."
What should you do this week?
You can do all of this in a 20-person business without a marketing agency or a new system.
- Gather last month's output. Pull together the posts, emails, proposals and web copy that went out in the last four weeks. Do not ask who used AI; the point is the words, not the author.
- Run the receipt check on three pieces. Highlight every claim, number, quote and promise, and write its receipt beside it. Anything without a receipt gets cut or fixed.
- Start a receipts folder. Customer quotes you have permission to use, before-and-after photos, short write-ups of finished jobs, your written policies and product sheets. This is what AI should draft from.
- Give AI one job for four weeks. Same-day follow-ups are usually the best first choice. Name one person who owns it.
- Add one line to your AI policy. "AI never writes a review, testimonial or customer quote." If you do not have a policy yet, a one-page AI policy is enough for most small firms.
- Share the standing instruction. Put "Use only the facts I have given you" at the top of every prompt anyone writes for sales or marketing.
Where to go next
If your team is already using AI to write marketing but nobody has shown them how to use it well, that is exactly what my AI training for teams is built for: practical sessions on your own sales and marketing jobs. If you want to see where your business stands first, take the free Business AI Readiness Scorecard: 12 questions across six areas, about five minutes.
Sources and further reading
- 2026 AI Impact Report: Mapping Adoption, Use, and Impact Across Small to Midsize Businesses, Intuit QuickBooks, May 12, 2026 (34,000+ survey responses from business owners in the US, Canada, the UK and Australia, with payment data from 5.3 million+ QuickBooks businesses).
- The AI Ad Gap Widens, Interactive Advertising Bureau, January 15, 2026 (505 US Gen Z and Millennial consumers and 104 US ad executives, October 2025 to January 2026).
- Federal Trade Commission Announces Final Rule Banning Fake Reviews and Testimonials, Federal Trade Commission, August 14, 2024.
- Fake reviews: Guidance on the prohibition under paragraph 13 of Schedule 20 to the Digital Markets, Competition and Consumers Act 2024 (CMA208), Competition and Markets Authority, April 4, 2025.
Dan Fitzpatrick is The AI Educator, a Forbes contributor and international keynote speaker who has worked with organizations including Google, Grammarly and Welsh Water, and who helps businesses use AI well, safely and every day.
