Quote & Proposal Generator
Slug: ai-quote-proposal-generator · Category: Sales & Quoting · Provisioning: Requires AgentLane staff approval
Drafts client quotes and proposals from a structured enquiry, with the arithmetic checked in a code step rather than left to the model. Read the limitation note below before selling it.
Required credentials
| Provider | Used for |
|---|---|
anthropic | Matches the enquiry to a service and drafts the quote copy |
postgres | Legacy, unused. No step in this flow reads or writes Postgres — see the note below. |
resend | Sends the quote email |
customjs | Not a stored credential — a step-capability flag for the inline code steps. There is no PDF generation. Doesn’t block provisioning on its own. |
Which AI actually runs this: AgentLane’s own managed Google Gemini models on Vertex AI. No Anthropic or OpenAI key is read for inference. The
anthropicrow above is a legacy credential requirement kept so existing rows and portal chips keep working — it is never checked before provisioning. See AI Models & Credits.
postgresis not used either. No catalog flow contains a Postgres step, so this agent writes nothing to a database of yours. The requirement is declared for historical reasons and is skipped before provisioning.
Industry fit
Home Services, Dental, Aesthetics & Beauty, Construction, Solar, Automotive, Insurance, Accounting, Legal — any business that prices jobs case-by-case rather than off a fixed price list.
How to deploy
Request it the same way as any non-instant agent — see Deploy your first agent for the request-dialog walkthrough.
Sample conversation flow
Webhook (quote request) ──► Look up client config ──► Draft proposal (Gemini)
│
▼
Parse proposal (code)
│
▼
Apply pricing rules (code)
│
▼
Send quote email (Resend)The model matches the enquiry to a service and drafts the quote copy; a code step then checks the arithmetic before anything is sent.
Known limitation — do not sell this agent on its pricing yet. There is no place to store per-client pricing rules, so the pricing step has no rule set to apply: it validates the model’s own arithmetic rather than replacing it with your numbers. In testing, three identical requests produced three different totals. Until a pricing-rules feature ships, treat every figure this agent produces as a draft for a human to price, and do not put it in front of an end customer unreviewed.
There is no PDF. The quote is sent as an email body; no step renders, attaches or encodes a PDF, and nothing is written to a database or posted to Slack.