saas-brand-intake

name

saas-brand-intake

description

Run SaaS brand intake interviews when the user wants to validate a new product idea. Use this when there is no existing brand directory and the user mentions a SaaS idea, says "I want to validate", "let's start a new brand", "I have an idea", "validate my SaaS", or runs `/new-brand`. Collects brand basics, SaaS context, and visual direction in three question clusters, then writes `brands/<slug>/brand.json` and `saas-context.json` and hands off to `saas-visual-generator`.

SaaS Brand Intake

You are conducting the brand intake interview — Step 1 of the SaaS marketing validation flow.

When to activate

• First chat in a repo where no brands/<slug>/ exists yet.
• User says any of: "I have a SaaS idea", "let's validate", "start a new brand", "new product", "help me launch", /new-brand <name>.
• User explicitly mentions they want a logo/landing page but there is no brand data yet — start here first.

What you must produce

Two files, both schema-valid:

• brands/<slug>/brand.json — conforms to templates/brand.schema.json
• brands/<slug>/saas-context.json — conforms to templates/saas-context.schema.json

How to interview

Use AskUserQuestion in clusters of 3–4 questions max. Never ask all 12 questions at once. Never ask one at a time either.

Cluster 1 — Brand Basics

• Brand / product name (derive slug from this, kebab-case, lowercase, ASCII)
• One-line pitch ("X for Y that does Z")
• Target audience: role + company size + industry
• Tone: 3 adjectives + 1 anti-adjective ("we are NOT __")

Cluster 2 — SaaS Context

• SaaS category (dev-tools, vertical-saas, ai-app, infra, analytics, productivity, …)
• Primary user job-to-be-done (one sentence)
• Pricing model intent (freemium / free-trial / usage-based / seat-based / enterprise / lifetime / tbd)
• Stage (pre-launch-waitlist / private-beta / public-beta / ga / post-launch)
• Distribution wedge (plg-self-serve / sales-led / community / content / open-source / marketplace)
• Competitors / inspirations (free text, ideally URLs)

Cluster 3 — Visual Direction

• Mood keywords (2–5, e.g. "calm, technical, optimistic")
• Color preference (warm / cool / mono / vivid / muted) + any hex anchors
• Logo type (wordmark / lettermark / icon-wordmark / abstract-mark / emblem)
• Brands they admire visually (free text)

After the interview

1. Write both JSON files. Use Read on the schemas first to confirm field shapes.
2. Reply in chat with a short recap: name, slug, ICP, tone trio, mood keywords.
3. End with: "Generating logo variants and moodboard now via Higgsfield."
4. Activate saas-visual-generator to run Step 2 automatically — do NOT wait for permission.

Delegation

For long or noisy intake (lots of free-text exploration), spawn agents/brand-intake.md. For a tight, focused intake, run it inline.

Edge cases

• If the user gives you everything in one paragraph, parse what you can and only ask clusters for missing required fields.
• If they refuse to answer something (e.g. competitors), set the field to [] or "tbd" and continue — don't block the flow.
• If they want to skip the visual direction cluster, default to: mood: ["modern", "clean"], colors.preference: "cool", logoType: "wordmark", and note it as defaults.
• Slug collision: if brands/<slug>/ already exists, ask whether to overwrite, create <slug>-2, or open the existing brand.

Related

• See questions.md for the exact AskUserQuestion payloads.
• See examples.md for sample completed brand.json files.
• Schemas: templates/brand.schema.json, templates/saas-context.schema.json.