Positioning an AI Product Without the Hype

By Brian Tr · Updated Oct 5, 2026

Positioning an AI product well means telling a buyer which job it does, for whom, compared with what they do today, and how far they can trust the result. It doesn't mean putting "AI-powered" in the headline. That phrase was a differentiator for a short while. It's now a claim every competitor makes, and, increasingly, a claim that regulators read closely.

This page applies established positioning frameworks to AI products. It covers why "AI" stopped working as a message, how to use April Dunford's components, jobs-to-be-done and Geoffrey Moore's positioning statement, why the competitive alternative is usually a person or a spreadsheet rather than other software, how to talk about accuracy and human oversight, and what the US Federal Trade Commission (FTC) and Securities and Exchange Commission (SEC) have said about overclaiming.

Why "AI-Powered" Stopped Being a Differentiator

A differentiator has to be true of you and not true of your alternatives. "Uses AI" fails both halves now. Organizational AI adoption reached 88 percent in 2025, according to the Stanford 2026 AI Index, and the same report counts 1,953 newly funded AI companies in the United States that year. When most buyers already use AI and thousands of startups claim it, the word tells a buyer nothing about why they should choose you.

There's a second problem. "AI-powered" describes how a product is built, and buyers don't buy construction methods. They buy a result. A finance director evaluating invoice software wants fewer manual keystrokes and fewer errors, and whether a model or a rules engine produces them is secondary. If your headline describes the technology, the buyer has to do the work of translating it into value, and most won't.

This is the practical consequence of the definition in what is an AI-native startup: the AI is the engine, but the outcome is the product. Positioning should follow the product, not the engine.

The Frameworks, Applied to AI

Four well-known frameworks handle most of the work. None of them was written for AI, and that's the point: they're about how buyers decide, and that hasn't changed.

Ries and Trout: occupy a position in the buyer's mind

Al Ries and Jack Trout's Positioning: The Battle for Your Mind frames positioning as the place a product holds in a prospect's mind relative to competitors, not something you do to the product itself. For AI products the lesson is about crowding. If ten vendors in your category say "AI-powered," that slot is full, and a buyer's mind has no room for an eleventh. The open positions are the specific ones: a named job, a named industry, a named result.

April Dunford: five components

April Dunford's positioning work, set out in her book Obviously Awesome, breaks positioning into components that build on each other. The commonly cited list is: competitive alternatives, unique attributes, value, target customer characteristics and market category, with relevant trends as an optional sixth. Here's what each looks like for an AI product.

The order matters. In a Gainsight interview, Dunford says that if you get the competitive alternative wrong, everything downstream is wrong. That's why the next section spends time on it.

Jobs to be done: name the job

The jobs-to-be-done view, laid out by Clayton Christensen and coauthors in the Harvard Business Review article Know Your Customers' "Jobs to Be Done", starts from the job a customer is trying to get done rather than from customer demographics or product features. Applied to positioning, it gives you a test: can you finish the sentence "People hire this product to..." without using the word "AI"? If the sentence only works with the technology in it, you haven't found the job yet.

A usable job is specific and has a "before" and "after." "Turn a pile of inbound supplier invoices into approved payments without retyping them" is a job. "Leverage AI for finance" is not. For more on mapping jobs, pains and gains to a product, see the value proposition canvas.

Geoffrey Moore: the positioning statement

Geoffrey Moore's positioning statement, from Crossing the Chasm, is a one-sentence template. It runs roughly: for [target customer] who [need or opportunity], our [product] is a [category] that [key benefit]; unlike [competitive alternative], our product [primary differentiation]. It's useful because it forces you to fill every blank, and AI products tend to skip two of them: the competitive alternative and the specific benefit.

Here's an illustrative template, not a real company:

For accounts payable teams at mid-sized distributors who spend days keying supplier invoices, [Product] is an invoice-processing service that turns invoices into coded, approval-ready entries. Unlike manual entry or templated OCR, [Product] handles new supplier layouts without setup, and routes uncertain fields to a person for review.

Notice that "AI" appears nowhere. The mechanism is implied by "handles new supplier layouts without setup," which is a unique attribute a buyer can verify. For a broader treatment of the concept, see competitive positioning and the value proposition page.

Choose the Right Competitive Alternative

Most AI founders name other AI startups as their competition. Buyers rarely do. A buyer's real alternatives, in rough order of how often they win, are:

1. Do nothing. The problem is annoying but tolerable, and the status quo costs nothing to keep.
2. A person. An employee, a contractor, an outsourced team or an intern does the work today.
3. A spreadsheet or a manual process. The work happens in tools the buyer already owns.
4. Existing software with a new AI feature. The incumbent platform they already pay for now includes something close enough.
5. Another AI-native product. This is the competitor founders worry about, and it's often the last one the buyer considers.

Dunford makes the same point in the interview above: the competitive alternatives to what you do often don't look like direct competitors, and the real competitor is often "do nothing" or "hire an intern to do it."

The choice changes the message. If your alternative is a person, your positioning is about throughput, consistency and cost per task, and your proof is a side-by-side with that person's output. If it's an incumbent's built-in feature, your positioning is about depth: the part of the job the built-in feature only half does and you finish. If it's another AI product, you need a narrower category or a specific data advantage, which is the territory covered in AI wrapper vs AI-native and vertical AI vs horizontal AI.

Lead with the Outcome, Prove with the Mechanism

Once the alternative is clear, structure the message in three layers.

The outcome, in the buyer's words. "Close the books three days faster." "Answer routine support tickets before the customer follows up." State the result a buyer can measure in their own business.

The mechanism, in one plain sentence. Explain how it works at the level a skeptical buyer needs. "It reads each invoice, proposes the coding, and flags anything it's unsure about for a person." This is where AI belongs, as the explanation of how, after the what.

The evidence, in numbers you can defend. Customer-measured results, test results on the buyer's own data, or a pilot. If you don't have them yet, say what you'll measure in a pilot rather than inventing a benchmark.

Be Explicit About Accuracy and Human Oversight

AI products fail differently from traditional software. They don't crash; they return a plausible wrong answer. Pretending otherwise is a positioning mistake as well as an ethical one, because the buyer who discovers the limits after purchase becomes a detractor.

Stating limits can itself be positioning. Three habits help:

• Say what the product is accurate at, and on what. An accuracy figure without a dataset, task and date is a marketing line. A figure with all three is a claim a buyer can check.
• Show where a human sits in the workflow. "Routes low-confidence results to a reviewer" tells a risk-averse buyer that someone is accountable. In regulated or high-stakes work, human oversight is often the feature that makes the purchase approvable at all.
• Say what it doesn't do. A one-line scope boundary ("it drafts; it doesn't give legal advice") removes a class of objections and a class of liability.

The accuracy question has teeth. In April 2025, the FTC announced a proposed order against Workado, which had promoted its AI content detector as "98 percent" accurate. According to the FTC's press release, independent testing showed the accuracy rate on general-purpose content was 53 percent, and the FTC alleged the underlying model had only been trained or fine-tuned to classify academic content. The proposed order barred Workado from making effectiveness claims about covered products without competent and reliable evidence at the time of the claim. The FTC approved the final order in August 2025. The lesson for positioning is narrow and practical: an accuracy number is a legal claim, so it needs a test behind it that matches the population the claim implies.

The Legal Risk of Overclaiming ("AI Washing")

"AI washing" means overstating, or inventing, the role AI plays in a product or business. Two US agencies have acted on it, and both statements are on the record.

The FTC. On September 25, 2024, the FTC announced five enforcement actions it described as a crackdown on deceptive AI claims. In the press release, then-Chair Lina Khan said: "there is no AI exemption from the laws on the books." One case involved DoNotPay, which the FTC said claimed to offer "the world's first robot lawyer." The FTC's complaint alleged the company had not tested whether its chatbot's output matched a human lawyer's and had not hired or retained any attorneys. DoNotPay agreed to a proposed order that required a payment of $193,000, a notice to subscribers about the limits of its law-related features, and a ban on claiming it can substitute for a professional service without evidence. Other actions in the same release concerned business-opportunity schemes that promised income from "AI-powered" storefronts.

The SEC. On March 18, 2024, the SEC announced settled charges against two investment advisers, Delphia (USA) Inc. and Global Predictions Inc., over false and misleading statements about their use of AI. In its press release, the SEC said Delphia's statements that it used AI and machine learning with client data in its investment process were false and misleading because it didn't have those capabilities, and that Global Predictions had falsely claimed to be the "first regulated AI financial advisor." The firms paid civil penalties of $225,000 and $175,000. The Director of the SEC's Enforcement Division also said that public issuers making claims about their AI adoption "must also remain vigilant" about similar misstatements.

These cases involved an FTC consumer-protection action and an SEC securities action, and the facts differ. What they share is the logic: if you say your product does something with AI, you need to be able to show it, at the time you say it. The same applies to a startup's pitch deck, website and press release, since investors and buyers both rely on them. This isn't legal advice, and the rules vary by country and industry, so a startup making performance or capability claims should have counsel review them.

A short pre-publication check:

Key Facts

• "AI" is no longer a differentiator: 88 percent of organizations reported AI adoption in 2025, per the Stanford 2026 AI Index, which also counts 1,953 newly funded US AI companies that year.
• Start from the alternative: April Dunford says that if you get the competitive alternative wrong, everything downstream is wrong, and that the real alternative is often "do nothing" or a person (Gainsight interview).
• Accuracy claims are legal claims: the FTC's Workado order followed a "98 percent" accuracy claim against 53 percent on general-purpose content in independent testing (FTC).
• There is no AI exemption: the FTC's September 25, 2024 sweep covered five companies, including DoNotPay, which agreed to a proposed order with a $193,000 payment (FTC).
• AI washing has securities consequences: the SEC's March 18, 2024 settlements with Delphia and Global Predictions totaled $400,000 in civil penalties (SEC).

How Positioning Connects to Go-to-Market

Positioning is upstream of everything a growth team does. The target customer decides where you advertise and who you call. The competitive alternative decides what you demonstrate and what you compare against. The category decides which analysts, directories and search terms you show up under. If you've nailed the first three components, the choices in your go-to-market framework get easier, and your pricing story from AI startup business models has something to anchor to: an outcome buyers already value.

If you're repositioning an existing product, the same sequence applies. Re-ask the five buyers what they do today, rewrite the statement with every blank filled, delete every instance of "AI-powered" you can't defend, and see whether the message still lands. If it doesn't, the problem was never the wording.

Related Reading

• Acquiring customers for AI products
• Why demos matter more for AI products
• What is an AI-native startup?
• AI wrapper vs AI-native
• Vertical AI vs horizontal AI
• AI startup business models
• Competitive positioning
• Value proposition canvas
• Go-to-market framework