> **PromptEye Brand Book & Design Language** · Idea and messaging · EN (translation of the Polish source; the Polish file wins on conflict) · state 28.10.2026
> For people: https://brand.prompteye.com/en/brand/ · Polish source: https://brand.prompteye.com/pl/marka.md · index for agents: https://brand.prompteye.com/llms.txt
> File paths are relative to https://brand.prompteye.com/ds/ (e.g. `tokens/colors.css` → https://brand.prompteye.com/ds/tokens/colors.css); `*.md` references resolve to https://brand.prompteye.com/ds/docs/en/ (translations) or https://brand.prompteye.com/ds/docs/ (Polish source). Outside Claude Design load `styles.css` and the scripts `js/pe-light.js`, `js/pe-pl.js`, `js/pe-raster.js`, `js/pe-motyw.js` (in this order) instead of `_ds_bundle.js`.

# Brand: idea, messaging, method

**Status: in force (27.10.2026).** All PromptEye communication is built on this idea: the website, use cases, case studies and the visual motif. Changes come from the team.

## Idea

**AI recommends your brand in the questions that bring customers.**

(Polish original: „AI poleca Twoją markę w pytaniach, które przynoszą klientów.”)

- **"Recommends", not "is visible".** Visibility is the category word; every competitor uses it. What the customer cares about is that AI points to their brand when someone asks what to buy or whom to choose. AI is becoming the most important recommender today.
- **"In the questions that bring customers", not "everywhere".** This sets us apart from tools that show a general visibility score. Being in the answer to a question that sells nothing means nothing.

Being recommended by AI where it makes sense builds brand awareness and, as a result, brings customers. The outcome is not the same for every customer, which is why the idea connects to use cases.

The same idea is the theme of the visual motif: the field of dots is the market of questions, the object emerging through light is the questions that bring customers, the orange point is the brand recommended by AI (`motyw.md`).

## Why we build on it

In the age of AI, features can be built or copied in a week. Communication based on features ages together with them. The idea has to be true today, in six months and in a year.

Three things cannot be copied (differentiators):

1. **The effect.** "AI recommends your brand" stays true whatever models and features exist a year from now.
2. **People and their track record.** Years in SEO, exits, GEO done for agency clients before the tool existed.
3. **Data collected over time.** Measurements of whole industries, reports, the Atlas, the history of model answers. A competitor can copy a feature but cannot measure backwards how AI talked about a market six months ago. This advantage grows every month.

## Order in communication

From the most durable to the fastest-changing:

1. The effect (permanent).
2. Why it works with us: method, people, data (permanent).
3. Use cases (change slowly).
4. Features (change every week). We show them only as proof that the method can be carried out, never as the main message.

## Use cases: forms of one effect

Different personas use PromptEye and work with us through use cases, i.e. concrete usage scenarios. We describe every use case as a variant of the same effect, so the communication stays consistent even though the goals differ.

| The persona wants to… | The effect in the language of the idea |
|---|---|
| Win customers | AI recommends the brand in buying questions |
| Shape how the brand is perceived | AI describes the brand the way it wants to be seen |
| Publish many good articles for SEO/GEO | The content is cited by AI in important questions |

This is the start of the list; there are more use cases. **Describe a use case by its effect, not by the activity.** "Publishing articles" is an activity in which we compete with every content generator. "Content cited by AI" is an effect that keeps us in our category. Do not add new use cases without the team.

## The method in three steps

Every use case answers two customer questions: how do we know AI will recommend the brand, and how do we know these particular questions bring customers. One method, shared by all use cases, answers them:

1. **We choose the questions that bring customers.** We know which ones thanks to SEO experience (purchase intent) and data from measuring whole industries.
   In practice: selecting questions and grouping them into topics and intents, assessing the quality of the question set, business priorities, suggestions on which questions to add, remove or replace.
2. **We change what decides the AI answer:** content, sources and publications, in one place. Buying links and publications directly in the system is a road to the goal, not a differentiator in itself.
   In practice: recommendations based on monitoring, AI bot traffic and technical analysis of the site; content gaps against competitors; creating and optimising content; buying links and publications in the system.
3. **We check whether AI has started recommending and whether customers came.**
   In practice: monitoring AI model answers (share of recommendations and position against competitors, how the brand is described, sources and citations), AI traffic on the customer's site, change in visibility after published content.

The same elements appear in the offer as layers: monitoring is steps 1 and 3, visibility management is step 2. **In communication use the steps of the method**; the layers stay in the offer and the price list. The ChatGPT app is a separate enterprise product, outside the core offer and outside the method.

The method has no name yet. **Do not invent one.** Describe it with the three steps.

## The team and the limit of promises

The "we have an experienced team" story on its own is weak, because every start-up says it. It works only when the team stands behind the method and case studies show that the method gives the same result in different situations.

To the question "how do we know these questions bring customers" we answer today with experience. A full answer from data requires measuring AI traffic and conversions at the customer. Until we have that, in promises about customers and conversions we say "we help", not "we do".

## Case study: fixed template

Every case study has the same structure. Repeatability shows when the "why it worked" section returns to the same three steps of the method every time.

1. Who and which scenario.
2. Starting point: a measurement from the tool.
3. What we did: the three steps of the method.
4. Result: numbers from the tool and, where possible, traffic or customers.
5. Why it worked: the same three reasons.

Only confirmed case studies. Where there is no customer yet for a use case, there are two honest substitutes: a case from the agency GEO days, or PromptEye as its own case (SEO/GEO on a new website).

## Vocabulary: rules

- **Yes:** recommends, recommendation, AI recommendation, questions that bring customers, effect, method, measurement.
- **Careful:** "visibility" only as the name of a metric in data ("visibility in the model, %"), never as a promise.
- **"We help", not "we do"** in promises about customers and conversions (section "The team and the limit of promises").
- **"We have an experienced team"** is not an argument on its own. The team always stands behind the method and the cases.
- **Features** only as proof that the method can be carried out, never as the main message.
- **Numbers:** only confirmed ones. Do not use "93 out of 100 pieces of content raise visibility in the first month" until it is confirmed.
- Polish spelling and typesetting rules: `typografia.md` (P1).