Quote / Proposal Skill

Fixes: the owner hand-building quotes, proposals, or estimates after hours. Turn 90 minutes of nightly Word-wrangling into a 2-minute review. Pure efficiency win, and one of the most felt — it gives the owner their evenings back.

Stack

• Brain: Claude skill (or Claude Project) with the client’s price book, service descriptions, terms, and brand voice as context files.
• Input: dictated or typed job notes (voice → text via phone, or pasted).
• Output: formatted, priced quote/proposal for review → send as PDF or link-shareable Google Doc (~/google/md_to_gdoc.py pattern).
• Optional: trigger from CRM/form so an inquiry auto-drafts a starting quote.

field-tested build

1. Audit — how are quotes built today? What varies vs. what’s boilerplate? Where do errors/inconsistencies creep in? How long does it take?
2. Optimize — standardize the price book and the quote structure first. Kill the parts that don’t need to be bespoke. A clean, consistent template is 80% of the value; the AI just fills it.
3. Automate — build the skill: Input (job notes) → Process (match to price book, apply standard terms, compute totals) → Output (formatted quote, owner reviews). Human approves every quote early on.
4. Verify — run 3 real past jobs through it; check the math, the scope language, and that it stayed within scope (no invented promises). Would the owner send it with minor edits?

Time to ship

Quick Win — ~1 session once the price book exists (gathering the price book may be the slow part).

Compliance

Financial/scope accuracy is on the human reviewer — the skill drafts, the owner signs. Note this explicitly in the workflow.

Log in Notion

“Built quote skill. Quote time [before, e.g. 90 min] → [after, e.g. 2 min review]. ~[hrs]/week back.” Attach a sample generated quote.