Example automation · Trades

Quotes drafted from the job request

A customer's job request and photos come in through the website. The automation drafts a line-item quote from your own price book for the estimator to check and send.

Try it: open an item, then approve it or take it over. The data is fictional sample data and this is an example of what we'd build, not a client project.

Quote DeskExample · sample data

Requests waiting0 of 3 handled

Dana K., Sherwood Park · Yesterday

Basement floor, vinyl plank

What arrived · Website form + 4 photos

"Want to redo the basement floor with that vinyl plank stuff. Its all carpet right now over concrete. Rooms are about 20 by 32 plus a little bathroom. Can you quote this week? thx Dana" [photo 1-4: carpet, wall with tape measure at 8 ft, stairs, utility door]

What it read

Job
Supply and install LVP97%
Area
~640 sq ft (20 x 32 ft; bathroom not included)88%
Existing floor
Carpet over concrete95%
Photos
4 attached, no damage seen90%
Wants
Quote this week99%

% is how sure the software is of each field. Under 85%, a person checks it.

Draft quote

Remove and dispose carpet, 640 sq ft
Underlay, 640 sq ft
LVP mid-grade, 640 sq ft + 10% waste
Install labour, 640 sq ft
Transitions x 3

All prices from price book v2026-10. Bathroom floor: ask Dana whether it is included.

What it could save

Estimate · not a client result

For a 15-person renovation and finishing company. These numbers are assumptions, not measurements: the audit measures yours.

By hand today
12 requests a week × ~30 min each
≈ 6 h a week
With the automation
12 requests a week × ~10 min to review each
≈ 2 h a week
Difference
6 h − 2 h
≈ 4 h a week back

By hand: read the request, call back, study photos, look up prices, build and send. After: the estimator checks lines and fixes the unpriced ones; the site-visit jobs are in the average.

The job today

  • The office reads the web form or email and calls the customer back for missing details.
  • The estimator looks at the photos and builds the quote line by line.
  • Prices are looked up in a spreadsheet or remembered.
  • Quotes go out days after the request, and some never go out at all.

How the automation runs

  1. Trigger

    1. Request comes in

    From the website form or an email. The customer, address and photos are saved to a new request.

    In: Website form · Outlook

  2. AI

    2. Read the job

    Job type, area, access and materials are read from the description and photos. Missing details are listed, not assumed.

    In: Photo + text AI

  3. AI

    3. Draft from the price book

    Each line is matched to a price book item. A line with no match is marked NOT PRICED, never given a made-up number.

    In: Google Sheets price book

  4. A person

    4. Estimator checks

    Adjust a line, approve, done. Unpriced lines and anything the photos raise are shown at the top.

  5. Your system

    5. Quote sent

    Created as a quote in your field-service app and emailed with your usual wording; the request is marked quoted.

    In: Jobber

Example tools: Website form · Jobber, Housecall Pro or QuickBooks · Your price book (spreadsheet). The audit confirms what you actually run.

What it checks

Before a person sees a draft, these run. A draft that fails one is held, with the reason shown.

  • Every priced line must match a price book item exactly. No match: the line is marked NOT PRICED.
  • Area read below 85% confidence, or in conflict with the photos: the estimator is asked to confirm it first.
  • A quote over a size you set (for example $15,000) always goes to the owner, not just the estimator.
  • Photos showing damage, rot, wiring or water: flagged for a site visit instead of priced.
  • Price book version is printed on the quote so a price can always be traced.

When something goes wrong

Edge cases are planned for, not discovered in production.

Customer gives no square footage
A short 'what we still need' email is drafted for the office to send; no quote is built on a guess.
Job is outside the trades or area you cover
A polite decline is drafted, and the request is logged so you can see what you turn away.
Existing customer already has an open quote
The draft links to it and asks whether this is a revision.
Jobber can't be reached when the estimator approves
The approved quote is held and retried; the estimator is told, not left wondering if it went out.

What lands, and where

Once a person approves:

Jobber
A client record and a quote with line items, ready to send.
Outlook or Gmail
The email to the customer, in your wording, with the quote link.
Google Sheets
A row on the request log: received, quoted, how long it took.

Where a person stays in charge

Every quote is checked by the estimator before it goes out. If the photos show something the price book doesn't cover, the draft says so instead of making up a price.

What it would cost

An Automation Build is from $8,000, with the price fixed before work starts. What moves it for this one:

  • How big and how tidy the price book is
  • Whether site photos are needed to quote
  • Which field-service app the quote is created in

All prices are in Canadian dollars. GST is added where it applies.

What the audit checks first

Before anything is built, the $2,500 audit measures the job in your business, including the numbers the estimate above only assumes:

  • How many requests a week, and how many never get a quote
  • How long from request to quote today
  • Which jobs are standard enough to draft and which always need a site visit

Have a job like this one?

The AI Automation Audit is $2,500, fixed: a working session on one workflow and a written plan in 5 business days, with costed options you keep. Credited in full if you build.

Request the $2,500 audit

Not ready to pay? Get a free video teardown of one workflow.

Describe your workflow

Don't see yours? Tell us about it. If it's repetitive and lives in software, we can likely automate it.

A person replies within one business day. No obligation.

We use what you send only to reply and to see which campaigns bring real projects. We never sell it. How we handle your data

More examples

A few of the automations we build, not the full list.

Request the audit · $2,500