Competitive reconnaissance with AI

Design systems and branding with AI · Lesson 2 / 20

Why study other people's brands

Not to copy, but to find the empty seat. Markets converge on two or three visual and verbal templates: every fintech is blue, every craft bakery uses kraft paper, every B2B tool promises efficiency gains. Your opening is the hole in that template.

Gathering material in thirty minutes

Pick eight to twelve competitors: three leaders, three peers, three niche players, and one or two from an adjacent industry. The adjacent ones matter more than people expect, because techniques that look ordinary there look fresh in your category.

Analysis prompt:
"Here is the homepage copy of a competitor [paste].
Fill in: 1) who it addresses (segment, price tier);
2) the single main promise, in their own words;
3) the proof they offer;
4) tone in three adjectives;
5) what they do NOT promise, though they could.
Answer only from the text. Do not infer."

Line five is the valuable one. What competitors stay silent about maps free positions more precisely than what they claim.

Turning notes into a map

Put the answers in a table and ask the model to find repetition: which words appear across six or more of ten sites, which colours dominate, which promise nobody makes. Words in that frequency core — innovative, reliable, team of professionals — go straight onto a stop list. They do not differentiate; they camouflage.

Two axes, not ten

A positioning map runs on two axes that matter to the buyer, not to you. Bad axis pair: quality versus price, because everyone places themselves top right. Working pairs: beginners versus professionals, fast and simple versus deep and configurable, personal craftsman versus industrial scale. The right axes come from how customers actually decide between options.

Insight. Read three-star reviews of competitors rather than their websites. Five-star reviews praise in generalities and one-star reviews vent, but three-star reviews describe a specific defect, which is a ready-made list of promises you can make instead of them.

Common mistake. Assembling the competitor list from your own mental model of the market. Customers choose from search results and recommendations, not from your list. Ask five real customers what else they considered and half the names will surprise you.

Pro tip. Tell the model to answer only from pasted text and paste the sources yourself. Asked about a competitor from memory, a model will cheerfully invent promises, numbers and a founding year.

Cheat sheet

Eight to twelve competitors, including one or two from an adjacent industry.
Track what they say and what they conspicuously avoid saying.
Frequency words from competitors go onto your stop list.
Map axes come from customer decision logic, not price versus quality.

1. Which analysis point contributes most to positioning?

2. Why does a quality-versus-price map fail?

3. What should you read instead of competitor marketing pages?

Task — checked by AI

Collect eight real competitors of your product, including one from an adjacent industry. Run their homepage copy through the analysis prompt from the lesson and build a table: segment, promise, proof, tone, silences. Submit the table plus three conclusions: which words turned out to be frequency noise (your stop list), which promise nobody is making, and the two axes you will use for your positioning map.

← Back

🔒 Answer the question correctly to move on to the next lesson.