Several bots under one billing account
Growth allows 3 chatbots; Business allows 10. You are not paying a new workspace tax for every landing page you shipped.
For agencies
Agencies already rewrite FAQs and hop between client inboxes. Point each bot at that client's site and PDFs, embed one script, and keep tenant data isolated. Growth includes 3 chatbots; Business includes 10 — one login for the studio.
Agency reality
Retainers cover design and campaigns. They rarely cover answering “what is your return window?” on ten different brands at 11pm. Clients still expect the site chatbot to know *their* policies, not a generic script you copy-paste.
Building a custom RAG stack per client kills margin. Sharing one knowledge base across clients is worse: it mixes answers and breaks trust. You need isolated bots with the same playbook.
How it works
Your account can hold multiple bots (plan-dependent). Each bot is a tenant: its documents, conversations, and widget key stay scoped to that client.
Crawl their public pages, upload SOPs and PDFs, add exact-match FAQs. Retrieval never searches another client's corpus.
Paste the widget script on their site. Connect WhatsApp or Instagram when their plan allows. Their agents take over with the full thread when the bot should not decide.
Built for studios
Growth allows 3 chatbots; Business allows 10. You are not paying a new workspace tax for every landing page you shipped.
Queries are tenant-scoped. Client A never answers from Client B's return policy.
Colours, greeting, and widget key are per chatbot so each site looks like the client's brand, not a template dump.
When the answer is not in the docs, the conversation goes to humans with the thread intact — your client's team, not a generic queue.
Volume and missed questions stay per client, so you can show value in the monthly report without mixing properties.
Remove our badge on Growth and above. Full white-label and SSO are Enterprise — we do not pretend otherwise on this page.
Compare
| Capability | Reply | One ChatGPT for all clients | Custom RAG per retainer |
|---|---|---|---|
| Tenant-isolated knowledge | YesYes | No | If you build it |
| Website widget + WhatsApp | YesYes | Not a site embed | Weeks each |
| Time to first client live | ~15 minutes | Prompt tinkering | Months |
| You operate the ML stack | We do | OpenAI does | YesYes |
FAQ