Playbook: Agent-first onboarding chatbot

Use this playbook when the human does not have a SiteGPT account, did not provide a SiteGPT token/profile, or asks to try SiteGPT from a website URL. The goal is to create value before signup: build, brand, train, and test a temporary chatbot, then share the onboarding URL where the human can preview and claim it. Do not require sitegpt whoami or sitegpt login before this flow. PROFILE_NOT_CONFIGURED is expected and is not a blocker.

sitegpt onboarding start https://example.com --agent-name "<agent-name>" --json

Capture:

Field	Use
`data.workspace.id`	Workspace ID for onboarding status, claim, and delete.
`data.workspace.chatbotId`	Chatbot ID for all setup commands.
`data.apiToken`	Temporary setup token scoped to this chatbot.
`data.onboardingUrl`	Human-facing preview and claim page.

For every setup command after onboarding start, pass the temporary token:

SITEGPT_API_TOKEN="<temporary-token>" sitegpt chatbots get <chatbot-id> --json

Do not print the full temporary token in final output.

2. Clarify purpose and inspect the website

Before configuring, know what job the chatbot should do. If the user did not say, ask:

What should this SiteGPT chatbot optimize for: customer support, marketing/site
guide, lead generation, docs/help, onboarding, or a mix?

If the user is unavailable and you should proceed, infer the purpose from the prompt and website, then state the assumption in the final report. Inspect the website with raw HTML for structured signals and browser/fetch tools for prose. Capture brand name, value proposition, important URLs, contact email, sitemap, tone, colors, icons, and product language.

3. Add knowledge for a strong first preview

Prefer sitemap ingestion. It usually gives the cleanest coverage.

SITEGPT_API_TOKEN="<temporary-token>" sitegpt knowledge sitemap add \
  --chatbot <chatbot-id> \
  https://example.com/sitemap.xml \
  --only-main-content true \
  --json

If no sitemap exists, crawl the website:

SITEGPT_API_TOKEN="<temporary-token>" sitegpt knowledge website add \
  --chatbot <chatbot-id> \
  https://example.com \
  --depth 3 \
  --max-links 100 \
  --only-main-content true \
  --json

If only specific pages matter, add selected links:

SITEGPT_API_TOKEN="<temporary-token>" sitegpt knowledge links add \
  --chatbot <chatbot-id> \
  https://example.com/pricing \
  https://example.com/docs \
  https://example.com/contact \
  --only-main-content true \
  --json

The preview should feel like the user’s website immediately. Use raw HTML/CSS/manifest signals for colors and icons. Download remote icons to local files before upload:

curl -sL https://example.com/logo-icon.png -o /tmp/sitegpt-logo.png
file /tmp/sitegpt-logo.png

Upload a clean icon when available:

SITEGPT_API_TOKEN="<temporary-token>" sitegpt icons upload --chatbot <chatbot-id> bot /tmp/sitegpt-logo.png
SITEGPT_API_TOKEN="<temporary-token>" sitegpt icons upload --chatbot <chatbot-id> chat-bubble /tmp/sitegpt-logo.png

Apply appearance:

SITEGPT_API_TOKEN="<temporary-token>" sitegpt settings appearance update \
  --chatbot <chatbot-id> \
  --title "<Brand> Support" \
  --welcome "Hi! I can help with <Brand> questions." \
  --placeholder "Ask about <Brand>..." \
  --brand-color "#0F766E" \
  --brand-text-color "#FFFFFF" \
  --link-color "#0F766E" \
  --icon-shape CIRCLE \
  --icon-position RIGHT

5. Create one strong persona

Create a single default persona that matches the purpose:

SITEGPT_API_TOKEN="<temporary-token>" sitegpt personas add \
  --chatbot <chatbot-id> \
  --title "<Brand> support specialist" \
  --file /tmp/sitegpt-persona.md \
  --json
SITEGPT_API_TOKEN="<temporary-token>" sitegpt personas use --chatbot <chatbot-id> <persona-id>

For customer support, emphasize accuracy, friendly escalation, and clear next steps. For marketing or lead generation, emphasize concise product explanation, qualification, and conversion-safe CTAs. For docs/help, emphasize grounded, source-based answers and careful unknowns.

6. Create one active instruction set

Instructions should cover grounding, uncertainty, escalation/contact behavior, and purpose-specific rules.

SITEGPT_API_TOKEN="<temporary-token>" sitegpt instructions add \
  --chatbot <chatbot-id> \
  --title "Grounded <Brand> assistant" \
  --file /tmp/sitegpt-instructions.md \
  --temperature 0.3 \
  --json
SITEGPT_API_TOKEN="<temporary-token>" sitegpt instructions use --chatbot <chatbot-id> <instruction-id>

Strong defaults:

Use SiteGPT knowledge first.
Do not invent prices, policies, legal terms, availability, or integrations.
Say when the knowledge does not contain the answer.
Offer the visible support/sales contact when relevant.
Keep answers concise and practical.
Ask a clarifying question when needed.

7. Add starters and followups for easy testing

Use prompts a visitor would naturally click.

SITEGPT_API_TOKEN="<temporary-token>" sitegpt starters add --chatbot <chatbot-id> --title "What does <Brand> do?" --message "What does <Brand> do?" --json
SITEGPT_API_TOKEN="<temporary-token>" sitegpt starters add --chatbot <chatbot-id> --title "Pricing" --message "Tell me about pricing and plans." --json
SITEGPT_API_TOKEN="<temporary-token>" sitegpt starters add --chatbot <chatbot-id> --title "Getting started" --message "How do I get started?" --json
SITEGPT_API_TOKEN="<temporary-token>" sitegpt followups add --chatbot <chatbot-id> --title "Contact support" --message "How can I contact support?" --json
SITEGPT_API_TOKEN="<temporary-token>" sitegpt followups add --chatbot <chatbot-id> --title "Compare options" --message "Can you help me choose the right option?" --json

For lead generation, include qualification and contact next steps. For docs/help, include docs, troubleshooting, and setup prompts.

8. Configure lead or support settings only when useful

If the purpose is lead generation and the site or user provides a sales email:

SITEGPT_API_TOKEN="<temporary-token>" sitegpt settings lead-form update \
  --chatbot <chatbot-id> \
  --enabled true \
  --collect-name true \
  --collect-email true \
  --notification-email sales@example.com

If the purpose is customer support and the site or user provides a support email:

SITEGPT_API_TOKEN="<temporary-token>" sitegpt settings general update \
  --chatbot <chatbot-id> \
  --support-email support@example.com

9. Verify with checklist and test messages

Check setup checklist:

SITEGPT_API_TOKEN="<temporary-token>" sitegpt onboarding status <workspace-id> --json

Fix PENDING, WARNING, or UNKNOWN checklist items when possible. Test realistic visitor questions:

SITEGPT_API_TOKEN="<temporary-token>" sitegpt messages send --chatbot <chatbot-id> "What does this company do?" --json
SITEGPT_API_TOKEN="<temporary-token>" sitegpt messages send --chatbot <chatbot-id> "How much does it cost?" --json
SITEGPT_API_TOKEN="<temporary-token>" sitegpt messages send --chatbot <chatbot-id> "How can I contact support?" --json

Share data.onboardingUrl only after the chatbot is useful. If the human wants to claim, ask for email, plan, and interval:

SITEGPT_API_TOKEN="<temporary-token>" sitegpt onboarding claim <workspace-id> \
  --email user@example.com \
  --plan GROWTH \
  --interval MONTH \
  --json

If data.checkoutUrl is present, give it to the human. If the email already has an active subscription, ask the human to open the onboarding URL while signed in and claim there. If the setup is wrong or the human does not want it:

SITEGPT_API_TOKEN="<temporary-token>" sitegpt onboarding delete <workspace-id> --yes

​1. Start without login

​2. Clarify purpose and inspect the website

​3. Add knowledge for a strong first preview

​4. Configure brand settings before sharing

​5. Create one strong persona

​6. Create one active instruction set

​7. Add starters and followups for easy testing

​8. Configure lead or support settings only when useful

​9. Verify with checklist and test messages

​10. Share and claim

1. Start without login