Yes. AI can draft a sound marketing plan for a small business in an hour, but only as good as the facts you give it: your goal, budget, customers, capacity and what has worked before. Without them you get a generic plan that would fit any business. Let AI supply the structure, options and calendar; you set the goal and the budget.
The usual failure isn't a wrong plan. It's a plausible plan nobody can carry out. Ask a chat assistant for a marketing plan and it will happily schedule daily social posts, a monthly newsletter, a podcast, webinars and paid ads for a business with one person and three spare hours a week. A plan you can't staff is worse than a short one you actually follow, so the most important input is the one owners leave out: how much time there really is.
What a plan needs that AI can't guess
Before any prompt, fill in an input pack. It takes about 30 minutes and it's the difference between a plan for your business and a plan for "a business like yours". An illustrative filled-in pack for a hearing-aid shop:
| Input | Example answer |
|---|---|
| One goal for the next 90 days, with a number | 20 new hearing tests booked a month, up from 12 |
| Monthly marketing budget | $600, fixed |
| Time available each week | Owner 3 hours, receptionist 2 hours |
| Who the customers are | Mostly over 65; many bookings made by adult children |
| How new customers find you now | Doctor referrals, word of mouth, Google search, a printed leaflet at two pharmacies |
| What has worked before | Leaflets at pharmacies; a free hearing check day at a community centre |
| What hasn't | Facebook ads aimed at over-60s: clicks, few bookings |
| Capacity limits | Two audiologists; maximum 30 new tests a month before waits grow |
| Rules and sensitivities | Health advertising rules; no claims about curing hearing loss |
| Seasonality | Quieter in summer; busy before the holidays when families visit |
The capacity line matters as much as the goal. Without it, AI can't know that doubling bookings would create a three-week wait, which would undo the goodwill the marketing built. If you're not sure who your customers really are, check that before you plan; how to check AI-generated customer personas explains how to test your assumptions against your records.
Set the budget from the value of a customer
"How much should I spend?" is the input owners most often leave blank, hoping AI will fill it in. It can't do so sensibly, because the answer depends on what a new customer is worth to you. The sum is short. An illustrative version for the same hearing-aid shop:
- About 40% of people who have a hearing test go on to buy hearing aids within three months.
- Average gross profit on a purchase, including the aftercare package, is about $1,500.
- So each booked test is worth roughly $600 in gross profit (40% of $1,500).
- The owner decides she's comfortable spending up to a quarter of that to win a test: $150.
- The goal is eight extra tests a month, so the ceiling is 8 × $150 = $1,200 a month.
She set the budget at $600, half the ceiling, for the first 90 days, so that a plan that underperforms still doesn't lose money. Now the AI has a real number to plan within, and the monthly review has a real test: is the cost per extra test below $150? Your own figures will differ, but the steps are the same for any business where you can estimate what a customer is worth.
A three-prompt sequence that produces a usable plan
One prompt asking for "a marketing plan" gets the generic version. Three prompts, each doing one job, get a plan you can use. They work in ChatGPT, Claude or Gemini.
Prompt 1: let it interview you.
I run a [business]. I want a 90-day marketing plan. Before writing
anything, interview me. Ask one question at a time, up to 12
questions, covering my goal, budget, weekly hours, customers, what
has and hasn't worked, capacity limits and any rules I must follow.
Wait for each answer. When you have enough, summarise my answers
in a table and ask me to confirm.Paste your input pack in answer to the first question and let it ask about the gaps. The confirmation table at the end is your check that it understood.
Prompt 2: draft within hard limits.
Now write the 90-day plan. Hard limits: no more than 3 channels,
total weekly effort within [X] hours, total spend within $[Y] a
month. For each activity give: what, who, hours per week, cost per
month, and the one number we'll check. No statistics or benchmarks
unless I supplied them. End with a month-by-month table.Prompt 3: stress-test it.
Act as a sceptical marketing adviser reviewing this plan for a small
[business]. List: the three activities most likely to fail and why;
any assumption not supported by what I told you; anything that
exceeds my hours or budget; and what you'd cut if the owner has a
bad month. Be specific.An illustrative extract of what prompt 3 returned for the hearing-aid shop's draft: "The community hearing-check morning depends on one venue and one audiologist being free; if either falls through, month 2 has no main activity. The plan assumes adult children read the leaflets, but you told me leaflets sit in pharmacies, where the older customer is more likely to see them. The weekly review replies are the cheapest activity and should be the last thing cut. If the owner has a bad month, keep the pharmacy leaflets and the review replies, and postpone the hearing-check morning rather than cancelling it." Every point in that answer was usable, and the second one led to a new line in the plan: a short version of the leaflet emailed to customers' family contacts, where the customer had agreed to it.
The third prompt is the one most people skip, and it catches the most. Models are much better at criticising a plan than at producing a realistic one first time, so make them do both.
Red-penning the first draft: what I'd cut
Even with good inputs, first drafts contain things to remove. An illustrative extract from a first draft for the hearing-aid shop, written before the hard limits were added:
Month 1: launch a TikTok channel sharing hearing health tips three times a week. Host a monthly webinar on hearing loss for adult children. Allocate $400 to local influencer partnerships. Goal: increase brand awareness by 30%. Email marketing typically returns many times its cost, so launch a weekly newsletter.
Taking it apart line by line:
- TikTok three times a week: not where the shop's customers or their families look for a hearing test, and far beyond five hours a week.
- Monthly webinar: a good idea in principle for adult children, but a large amount of preparation for a two-audiologist shop. Replaced with one printed guide for families, given out at appointments.
- $400 on influencers: two-thirds of the budget on a channel with no track record here. Cut.
- "Increase brand awareness by 30%": unmeasurable for a small shop. Replaced with the real goal: booked tests.
- The email return claim: a stock statistic with no source and no relevance to this shop's list of 400 customers. Cut, along with the weekly frequency. Monthly is plenty.
After the edits the plan had three channels: a monthly free hearing-check morning at a community centre (the thing that had worked), refreshed leaflets at four pharmacies instead of two, and a Google Business Profile post and review-reply routine every week. It fitted in five hours and $550 a month. The AI's structure and calendar survived; most of its channel ideas didn't.
A veterinary practice's 90-day plan, after edits
An illustration with round numbers. A three-vet practice has two goals: fill its nurse clinics (currently 55% booked) and sign up 60 more pets to its monthly health plan. Budget: $450 a month. Time: the practice manager has four hours a week, a nurse one hour. After the three prompts and a round of edits, the plan looked like this:
| Month | Activities | Owner and hours a week | Cost a month | Number to check |
|---|---|---|---|---|
| 1 | Health plan leaflet and counter card; reminder emails offering a nurse clinic to pets due a weight or dental check; weekly Google profile post | Manager 3h, nurse 1h | $180 printing and email tool | Nurse clinic bookings; plan sign-ups |
| 2 | Add a "new puppy and kitten" email sequence that introduces the plan; ask every plan member for a review | Manager 4h | $120 | Sign-ups from the sequence |
| 3 | Small local ad test ($250) promoting puppy and kitten health checks; review what worked and cut the rest | Manager 3h | $370 | Cost per new registration |
Illustrative outcome at day 90: nurse clinics at 78% booked, 41 new plan sign-ups (short of 60, mostly because the month 3 ad test brought registrations but few plan sign-ups), total spend $670 against a budget of $1,350. The review prompt below turned those results into the next quarter's plan: keep the reminder emails and the puppy sequence, drop the ad test, try asking at the end of every vaccination appointment instead. AI's time saving was in the drafting and the monthly reporting. The decision to drop the ads came from the practice manager looking at the numbers.
Turning the plan into weekly work
Plans fail in week three, when the busy season arrives and nobody remembers what Tuesday's task was. Ask the AI to turn the plan into a weekly checklist with named owners, and put the tasks in the calendar the team already uses:
Turn this 90-day plan into a week-by-week checklist. For each week:
the tasks, who does each one, how long it should take, and what
"done" looks like. Flag any week where the total exceeds [X] hours.
Format as a table I can paste into a spreadsheet.An illustrative week from the vet practice's checklist: "Week 5. Manager: send nurse clinic reminder email to pets due a dental check (30 min; done when sent and bookings logged). Manager: Google profile post, this week's topic 'hot weather and dogs' (20 min). Nurse: reply to this week's reviews (15 min). Manager: log plan sign-ups in the tracker (10 min)." Specific, small and checkable. If you use ChatGPT, its Work mode can produce the spreadsheet directly; otherwise paste the table into your own. For the content side of the plan, planning a month of marketing content with AI in one hour breaks posts and emails down further.
Signs an AI marketing plan is generic
- It would read the same with your business name swapped for a competitor's.
- It recommends every channel: social, email, video, podcast, ads, events.
- Goals are about "awareness", "engagement" or "presence" rather than bookings, sales or enquiries.
- It quotes statistics you didn't supply, with no source.
- No activity has a named owner or an hours estimate.
- The budget isn't broken down by month, or it exceeds the figure you gave.
- Nothing in it reflects what has already worked for you.
- It ignores your capacity: the plan's success would create a waiting list.
Three or more and it's a template. Go back to the input pack and the hard limits.
If you're new and have no history to feed it
A business in its first year can't fill in "what has worked before", and that's the line that keeps AI plans grounded. The fix is to make the first 90 days a test plan rather than a growth plan. Tell the AI you have no track record and ask for three small, cheap experiments on different channels, each with a four-week run, a spend limit and one number to judge it by. An illustrative new podiatry clinic ran a pharmacy leaflet, a Google profile with weekly posts, and letters to four doctors' practices, at under $200 a month in total. By week 12 the letters had produced most of the new patients, and the second quarter's plan was built around them. The AI wrote a sensible test plan; it would have written a fantasy growth plan if asked for one.
Which AI tools are enough for this
You don't need specialist software. Any of the main chat assistants on an individual plan will do: ChatGPT Plus and Claude Pro are $20 a month, and Google AI Pro, which includes Gemini, is $19.99. Two features help:
- Projects (in ChatGPT and Claude) or Gems (which become skills from November 2026) (in Gemini) keep the input pack, the plan and each month's results together, so you're not re-explaining the business every time.
- Research modes can survey what competitors publish, which is useful context for the plan. Treat every fact they return as something to check; doing competitor research with AI in one afternoon covers how.
Before you paste in customer lists or sales figures, remove names and contact details. A plan needs counts and trends, not personal data.
Reviewing the plan each month
A plan is only a plan if you check it. Once a month, paste the numbers into the same project and ask:
Here are this month's results against the plan: [numbers]. For each
activity: on track, behind or ahead, and the likeliest reason based
only on these numbers. Recommend one thing to stop, one to continue
and one small change to test next month. Keep within my hours and
budget.An illustrative monthly review for a dental practice running a plan to fill Saturday appointments: the AI reported that reminder texts had filled 70% of Saturday slots, the social posts had produced no bookings it could trace, and a referral card for existing patients had brought six new patients. Its suggestion: stop the social posts, keep the texts, print more referral cards. Sensible, and the owner took two of the three; she kept a weekly post because new patients said they'd looked at the page before booking, which the AI couldn't see in the numbers. That's the right split of work. Measuring whether AI is improving your marketing results covers which numbers to record so the review has something to work with.
Three businesses, three different plans
The same prompts produce very different plans when the inputs change, which is the best evidence the plan is really yours:
A pharmacy with a goal of more private consultations got a plan built around the calendar: vaccination season in autumn, hay fever in spring, travel health before holiday periods. Its main channels were in-store signage, prescription bag inserts and the Google profile, because its customers were already walking through the door.
An osteopath whose goal was steady new patients got a plan built on referral partnerships: two gyms, a running club and a local employer's wellbeing scheme, plus a short monthly email to past patients. Social media barely featured, because the input pack showed almost nobody had found the clinic that way.
A podiatrist whose goal was more diabetic foot checks got a plan aimed at referrers and families: letters to doctors' practices, large-print leaflets, and a line in reminder messages inviting family members to book for relatives. The AI's first draft had suggested Instagram Reels; the capacity and customer lines in the input pack are what removed them.
In each case, AI saved several hours of drafting and gave the plan a clear structure. The choices that made each plan work came from the owner's own facts. If you'd like a shorter version for the whole business rather than just marketing, writing a one-page AI strategy uses the same approach on a single page.
Marketing plans and AI: follow-up questions
Is it worth paying for a dedicated AI marketing plan generator?
Rarely. Most plan generators are a form in front of the same kind of language model you can use in ChatGPT, Claude or Gemini, with less control over the prompts. A general chat assistant plus a good input pack and the three prompts in this tutorial produces a better plan. Spend the money on the plan's activities instead.
How long should a small business marketing plan be?
Short enough to read in five minutes: usually one or two pages covering the goal, the customers you're targeting, two or three channels, a monthly budget, who does what each week, and the numbers you'll check. Long plans with a dozen channels tend not to get followed. If AI gives you ten pages, ask it to cut to the essentials.
Can AI decide how much I should spend on marketing?
It can show you common ways of setting a budget and do the arithmetic, but the figure depends on your margins, cash flow and how many new customers you can actually serve. Set the budget yourself from what a new customer is worth to you and what you can afford to test for three months, then let AI split it across the plan.
Further reads
- 12 AI Marketing Mistakes Small Businesses Make (and the Fixes) — The mistakes that turn a sensible plan into wasted spend.
- How Small Businesses Actually Use AI in Marketing: 15 Examples — Examples of AI tasks you might build into the plan.
- Can a One-Person Business Automate Its Marketing With AI? — Plan marketing when the owner is the whole team.
- Set Up Conversion Tracking Before You Let AI Spend Your Ad Budget — Set up tracking before any paid ads in the plan go live.
- Which Marketing Tasks Should a Small Business Never Hand to AI? — Which parts of the plan should always stay with a person.
- How to Fact-Check AI Marketing Copy Before It Goes Live — Check the copy that comes out of the plan's activities.
- How to Create Customer Personas With AI From Real Data — Build personas from booking, order and review data instead of AI guesswork: prompts, a persona template with evidence, and four businesses' results.
- How to Test a New Service Idea With AI Before You Launch It — A six-week way to test a new service with AI doing the drafting and real customers supplying the evidence, with pass marks set before the results arrive.
- AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.
Sources: ChatGPT, Claude and Google AI plan pricing pages; OpenAI's descriptions of Projects and ChatGPT Work.