Ask an AI tool to write a marketing plan for almost any business and you'll get something structurally sound and instantly forgettable: a target audience described in demographic generalities, three or four channel recommendations that would apply equally to a competitor, and messaging built around adjectives like "innovative" and "customer-focused" that say nothing a reader couldn't have guessed before opening the document. This isn't a limitation of the AI's writing ability — the structure and prose are often genuinely well-formed. It's a limitation of what it was given to work with. A marketing plan is only as differentiated as the positioning input behind it, and when that input is thin or absent, the AI does exactly what it's built to do: produce the statistically most likely, most generic version of a marketing plan for a business roughly like yours, because that's what "a marketing plan" looks like on average across everything it's seen.

The fix isn't avoiding AI in marketing planning — it's being deliberate about which parts of the work AI should accelerate and which part has to come from you first, before AI ever touches the document. The parts that make a plan actually yours — your specific positioning, your actual competitive differentiation, the real tradeoffs your customers weigh — have to be supplied as input, not generated as output. Once that input exists, AI becomes genuinely useful for structuring, drafting, and speeding up everything downstream of it.

Why the generic-plan problem is a garbage-in problem, not an AI-quality problem

When someone prompts an AI tool with "write a marketing plan for a boutique project management consultancy," the tool has nothing specific to differentiate that consultancy from the hundreds of similar ones implicit in its training. It fills the gap with the most common, safest pattern for that category — because from a purely statistical standpoint, that's the correct thing to produce given no more specific information. The plan that comes back sounds like it could have come back for any of a hundred similar consultancies, because functionally, it did.

This is the same underlying idea as retrieval-augmented generation applied to strategy work rather than customer support: a model without your specific facts will confidently generate plausible-sounding generic content in their place, and the fix is the same in both cases — hand it your actual, specific information rather than asking it to infer or invent what it doesn't have.

The input that actually differentiates a plan: a real positioning brief

Before opening any AI tool, the work that has to happen first is writing a short, specific positioning brief — not a mission statement, not a vague sense of "who we serve," but concrete answers to a small number of pointed questions:

  • Who specifically is this for, and who is it explicitly not for? Not "small businesses" but something with enough specificity to exclude someone — "operations leaders at 20-to-80-person logistics companies who've outgrown spreadsheet-based routing but aren't ready for enterprise software" excludes far more than it includes, and that's what makes it usable.
  • What does the customer currently do instead, and why does that fail them? Every product or service competes against a real current alternative, even if that alternative is "nothing, they just live with the problem." Naming that alternative specifically, and its specific shortcoming, is what messaging actually needs to address.
  • What's the one claim a competitor in your space genuinely couldn't make with a straight face? This is the differentiation test — if a rival could swap their name into your positioning statement without it sounding false, the statement isn't differentiated yet, regardless of how nice it sounds.
  • What tradeoff does your ideal customer knowingly accept to work with you? Real positioning usually involves giving something up — being more expensive, more specialized, slower to onboard, less flexible — in exchange for something the customer values more. A plan with no acknowledged tradeoff is usually a plan trying to be everything to everyone, which reliably produces generic messaging.

A worked example: positioning a niche B2B SaaS tool

Take a small SaaS company selling scheduling software for physical therapy clinics — a genuinely crowded category with several larger, well-funded competitors. A generic AI prompt about "marketing a healthcare scheduling SaaS" will default to messaging around "streamline your practice" and "save time," which is true of every competitor in the category and differentiates nothing.

Working through the positioning questions first surfaces something much more specific: this company's actual customers are solo and two-provider clinics, not larger multi-location practices, because the product deliberately doesn't build the multi-location administrative features the bigger competitors compete on — a real, explicit tradeoff. The current alternative for most of these small clinics isn't a competing product; it's a paper calendar or a generic scheduling tool never built for healthcare's specific needs (insurance verification reminders, treatment-plan-linked recurring appointments), and that mismatch is a specific, nameable failure rather than a vague inefficiency. The claim a larger competitor genuinely can't make with a straight face: "built only for practices under three providers, so you're never paying for administrative features you don't need and never will."

Once that positioning brief exists — four or five sentences, not a full document — it becomes the actual input for everything AI helps build next: audience-specific messaging that names "solo and two-provider PT clinics" instead of "healthcare providers," content ideas that address the paper-calendar-to-software transition specifically rather than generic "time-saving" claims, and channel recommendations that make sense for reaching genuinely small practice owners (which looks different than reaching hospital administrators, a mistake a generic plan for "healthcare scheduling software" would be prone to making by defaulting to enterprise-healthcare-marketing patterns).

Tip

If you can swap your positioning brief's language into a direct competitor's marketing and it would still sound believable, the brief isn't specific enough yet. Keep pushing on the "what tradeoff do we accept" and "what couldn't a competitor honestly claim" questions until the answer stops being portable to anyone else in your category.

Where AI genuinely speeds up the plan once the input exists

With a real positioning brief in hand, AI becomes legitimately useful across most of the plan's remaining structure, and this is where the actual time savings live. Drafting audience personas that translate the positioning brief into specific messaging angles for each segment; generating a first-pass content calendar structured around the real customer pain points named in the brief rather than generic industry topics; drafting channel-specific messaging variants (a LinkedIn post, an email subject line, an ad headline) that all trace back to the same core positioning rather than each reinventing the value proposition slightly differently; and building out a competitive-comparison framework that's honest about the tradeoff rather than pretending the product wins on every dimension — all of these are exactly the kind of structuring and drafting work AI handles well, because the differentiating judgment call (the positioning itself) has already been made by a human and supplied as input, and the remaining work is genuinely mechanical translation of that input into various formats.

What to keep out of AI's hands, and why

Two parts of a marketing plan should stay human-led even after a strong positioning brief exists. First, the actual pricing and packaging strategy — this depends on cost structure, competitive dynamics, and business goals that are rarely fully captured in any brief, and getting it wrong has direct revenue consequences, not just a slightly generic-sounding paragraph. Second, any claim about outcomes or results — "clinics using our software reduce no-shows by 30%" — should never be AI-generated language dressed up as a fact; if a number like that exists, it comes from real data, and if it doesn't exist yet, the plan should say so honestly rather than let a fluent-sounding draft imply a data point nobody actually verified.

Reviewing an AI-assisted draft for generic drift

Even with a strong positioning brief as input, individual sections of an AI-drafted plan can drift back toward generic language over a long document, especially in sections further from where the brief was originally referenced. A practical review step: read the audience description, the key messaging pillars, and the competitive differentiation section, and for each one ask the same swap-test question from the positioning brief itself — would this sentence still make sense with a competitor's name substituted in? Any sentence that passes that test (meaning it would still sound true either way) needs to be rewritten with something more specific from the original brief, not softened language that sounds nice but says nothing new.

Common mistakes in AI-assisted marketing planning

The most common mistake is skipping the positioning brief entirely and going straight to prompting for the plan, on the theory that AI will "figure out" the differentiation from a short description of the business — it won't, because differentiation isn't something to infer from a business description, it's a strategic choice that has to be made and stated before anyone, human or AI, can build a plan around it. A second mistake is writing the positioning brief but never actually feeding it into the AI drafting process explicitly — pasting it in once at the start of a long conversation and then drifting into unrelated prompts loses that grounding by the time you reach the channel strategy section several prompts later; re-anchoring to the brief explicitly at each major section keeps the output tied to it. A third is treating the finished AI-assisted draft as final without the swap-test review pass, which is exactly the step that catches generic language creeping back in over a long document.

The plan is only as good as the positioning underneath it

None of this is really a lesson about AI and marketing plans specifically — it's the same lesson that applies to nearly every use of AI in strategic work: the technology accelerates structuring, drafting, and iteration dramatically, but it cannot supply the one input that was always the actual hard part of the exercise, which is a genuine, specific, defensible point of view about who you serve and why you're different. Do that work first, on paper, in a few sentences, and AI turns a multi-week planning process into a multi-day one. Skip it, and AI just produces a well-formatted version of the same generic plan a template would have produced anyway.

The AI-Powered Marketing and Sales Engine walks through building a positioning brief like the one above in much more depth, along with more worked examples across different industries showing how the same brief translates into messaging, content, and channel decisions. The Marketing Plan Template Suite gives you a structure to build the full plan around once your positioning brief exists, Marketing Campaign Management is where that plan turns into scheduled, trackable execution rather than a document that sits unused after the planning meeting, and Digital Marketing rounds out the channel-specific tactics — so the speed AI adds to the process compounds a real strategic decision instead of papering over the absence of one.