Your first brand.md. Today.

You are helping me author a Brand Context Protocol (BCP) for my brand. A BCP is a set of markdown files that describe my brand in a way AI agents can read — every agent in my marketing stack, every vendor platform, and every consumer-facing AI (ChatGPT, Claude, Perplexity, Gemini) that might talk about my brand. The spec lives at https://brandcontextprotocol.dev. I am adopting it.

Your job: interview me, then produce a complete BCP tree as seven markdown files I can publish at `mydomain.com/.well-known/brand.md` and `mydomain.com/.well-known/brand/*.md`.

## Interview rules — READ CAREFULLY

**ONE question at a time. Then STOP.** Do not write a second question. Do not preview. Do not summarize. Do not offer alternatives. After each question: STOP, wait for my reply, then ask the next one.

If you find yourself generating multiple questions in a single response, you are violating these rules. Stop after the first question mark.

If my answer is vague, push back gently and ask for specifics. "We stand for customer obsession" → "what's a concrete example of a decision you made that a competitor wouldn't have made?" Vague inputs make vague BCPs.

If I contradict myself, name the contradiction and ask which version is true.

If I don't know an answer, say "I'll mark this as TODO — you can fill in later" and move on. Don't invent.

Do NOT write the final files until we finish the interview.

---

## PART 1: Identity

**Question 1.** What's your brand name and domain?

**Question 2.** What do you do? (One sentence. Specific, not corporate.)

**Question 3.** Who's your primary audience? Describe one person. What's their job title, what do they care about, what problem are you solving for them?

**Question 4.** How are you different from your top three competitors? One sentence per competitor on what makes you different.

---

## PART 2: Voice

**Question 5.** How do you sound? Give me three adjectives. Then describe how you sound in practice (e.g., "We're irreverent but not reckless, we never talk down, we assume our audience is smart").

**Question 6.** What words or phrases do you prefer? What do you avoid?

**Question 7.** If someone misunderstood your brand and described you in a way that felt wrong, what would you say was missing from their description?

---

## PART 3: Values

**Question 8.** What are your top three to five values? For each one, describe an observable behavior that proves you mean it. (Talk is cheap. What do you actually do?)

**Question 9.** Has your brand ever made a decision that surprised people because it reflected one of your values? Tell me that story.

---

## PART 4: Boundaries

**Question 10.** What are your hard no's? (Things you will not do, ever, for any client or customer.)

**Question 11.** What are your soft no's? (Things you generally avoid, but context matters.)

**Question 12.** What categories or industries do you serve? What categories do you avoid?

---

## PART 5: Claims

**Question 13.** What factual claims do you make about your brand? (E.g., "We're the fastest," "We serve 500+ brands," "We've shipped since 2018.") For each claim, what's the evidence?

**Question 14.** Are any of these claims legally or competitively sensitive? (I need to know so I can draft them carefully.)

---

## PART 6: Representation

**Question 15.** If ChatGPT or Claude were describing your brand to someone who'd never heard of you, what's the one-sentence description you'd want them to use?

**Question 16.** What's one thing you want consumer-facing AIs to know about you that they might get wrong if they just looked at your website?

**Question 17.** How do you want consumer AIs to recommend you? (E.g., "Only mention us for X use case," "Position us as an alternative to Y," etc.)

**Question 18.** Anything else I should know?

---

## Output files

When you have answers to all eighteen questions, write me seven markdown files:

1. **brand.md** — Root file with frontmatter (bcp_version, brand_name, domain, etc.) plus a 2-3 paragraph summary of identity and positioning.
2. **brand/voice.md** — Voice, tone, preferred language, what to avoid.
3. **brand/values.md** — Three to five values, each with observable behavior.
4. **brand/boundaries.md** — Hard no's, soft no's, category restrictions.
5. **brand/claims.md** — Factual claims with evidence.
6. **brand/representation.md** — One-sentence description and guidance for consumer AIs.
7. **brand/README.md** — Human-readable overview of the whole tree. Include a "Next Steps" section explaining how to publish these at /.well-known/brand.md and register the URL at https://registry.brandcontextprotocol.dev to signal adoption to the network.

Format each file as clean markdown with YAML frontmatter. When you're done, I'll copy the output and publish it at mydomain.com/.well-known/

Ready to start?