← All use cases

Use case B32 · Business

Draft and keep client proposals straight with AI, from the sales call, the estimate or the latest meeting: two remodelers and a landscape owner

Vlatko Demrovski, owner and GM, Element Design Build; Lee Mokri, Geared for Growing; and Ashley Sheaffer, principal, KBF Design Gallery ·

The short version

What they did

Three trades owners have AI turn what they already have, a recorded sales call, a finished estimate or a meeting transcript, into a client proposal or an update to one. Each checks the result before it goes out.

A remodeler’s proposals from recorded calls

Vlatko Demrovski records prospect calls and consultations with Fireflies. Claude extracts the discussion and action items, then helps prepare JobTread proposals from templates and historical company data. A separate Claude project assembles final scope and budget from selections and stored costs. He reviews the output. Qualified Remodeler, 2026-10-08.

A landscape owner’s proposals from the estimate

Lee Mokri gives an AI assistant his estimator’s spreadsheet, a proposal template and past proposals. It rewrites the scope for clients, formats the document and flags inconsistencies. He checks numbers and commitments before sending. The article separates this current workflow from his proposed future company-wide knowledge system. His column in Green Industry Pros, 2026-10-15.

A remodeler keeping selections and meeting decisions in step

Ashley Sheaffer combines the firm’s sales playbook, contracts, client communications and project data in Claude Cowork. A contract tool distributes material selections across proposal sections; uploaded Otter transcripts flag decisions made and still pending. It addresses changes that previously had to be reflected in six places. Custom Builder, 2026-05-20.

Result (as claimed)

The remodeler: the article says the initial proposal task “once took 60 to 90 minutes to prepare” and is completed “in 10 to 15” minutes. These are Demrovski’s reported times, not library measurements. The article also discusses business changes beyond AI, so overall growth is not treated as an AI-only result.

The landscape owner: “But it also misses things, like two different prices in the same proposal.” Mokri describes catching a $30,000 pricing table alongside a $28,000 maximum elsewhere. He reports quicker proposal turnaround qualitatively, without a measured after-time.

The selections and decisions: on the broader sales-intelligence workflow, Sheaffer says it helped her “take a step back out of the day-to-day sales management and more into a leadership role.” No measured savings or error-reduction rate is reported for the contract tool.

Prompt

From a recorded call:

I run a remodeling business and need an initial proposal from a consultation. The transcript is [attach], our approved template is [attach], and relevant past jobs and costs are [attach]. The prospect’s confirmed selections are [paste or say pending].

1. Extract the requested work and action items, citing the speaker and passage for each requirement.
2. Draft the initial scope using our template. Draw any costs only from the supplied job history and identify the comparable job.
3. Separate confirmed decisions from assumptions and list missing selections or prices as open items.

Use plain, professional wording and match the template’s length. Don’t invent commitments, dates or costs. Return a draft and short action list for my review, without sending it.

From a finished estimate:

I run a landscape business and need to turn an approved estimate into a client proposal. The estimate is [attach], our proposal template is [attach], and examples of our usual wording are [attach past proposals]. Help me keep the scope and numbers consistent.

1. Rewrite the scope in plain client language using our template. Preserve every approved price, quantity, date and exclusion.
2. Produce a separate consistency table comparing repeated values and commitments across the proposal and estimate. Identify each location.
3. Flag conflicting amounts, unclear scope and missing details for my decision, leaving them unresolved.

Match the template’s format and length. Don’t pick one conflicting price yourself or add promises. Return the draft, table and a short questions list. Do not send the proposal.

From the latest meeting:

I run a remodeling firm and need to keep a proposal aligned with meeting decisions and material selections. The latest transcript is [attach], the current selections are [attach], and the current proposal is [attach]. Our rules for confirming selections are [paste].

1. Compare the transcript with the existing documents. List confirmed changes, pending decisions and every section each change affects.
2. Cite the speaker and passage supporting each proposed change. Distinguish a discussed option from a confirmed selection.
3. Return a proposed update table: current wording or selection, proposed change, affected locations, evidence and open question.

Use plain language and keep the summary under 200 words. Don’t infer approval from silence or invent selections, prices or dates. Wait for my approval before changing the proposal.

Do it yourself

These steps are my suggested way in, not any one person’s setup.

Level: Beginner for one proposal in a chat. Linking call recordings and job software is Intermediate.

What you need

  • A chat assistant such as Claude or ChatGPT. A project feature that holds files helps.
  • Your proposal template and two or three past proposals you liked.
  • The source for the new one: an estimate, a call transcript or meeting notes.

Start simple

  1. Upload your template, past proposals and one finished estimate.
  2. Paste the “From a finished estimate” prompt above.
  3. Read the consistency table first. Resolve each conflict yourself.
  4. Edit the scope wording, then compare every number with the estimate.
  5. Send it yourself once it matches.

Take it further

  • Record sales calls with a note-taking tool and draft the first scope from the transcript.
  • After each client meeting, use the third prompt to list what changed and what is still open.
  • Keep one project per client with the proposal, selections and meeting notes together.

Check your work

  • Check every price, quantity and date against the estimate. One owner here caught two different prices in one proposal.
  • Ask it to list anything it assumed. Remove or confirm each item.
  • A proposal is a promise. Read the exclusions and commitments before it goes out, and get consent before recording a call.

Source checked. The link, author, date and quote were confirmed against the source. The result itself has not been reproduced. Prompt adapted for this library, inspired by the case. Business size: 8 employees; midsize landscape company (headcount not stated); remodeling firm (size not stated). Results are quoted as the source claimed them and never improved. Spot an error or a broken link? Email charlie@cbarmorecpa.com.

More in Sales

New cases go out every Saturday in Inside AI.