Discover
We look at recent quotes, the notes behind them, your price list, and how they are checked and sent today.
EXAMPLE PROJECT
AI that drafts quotes and proposals
A quote often means piecing together notes, a price list, and an old proposal. AI can assemble a first draft so a person spends their time checking it.
01 · THE PROBLEM
The problem
After a site visit or call, someone turns rough notes into a written quote, looking up prices and copying sections from earlier proposals. It often waits until the end of a long day.
Slow quotes can cost work, and copied sections sometimes carry over details from the wrong job.
02 · SIGNS
Signs you have it
- Quotes and proposals follow a familiar structure.
- Prices come from a list or spreadsheet that is kept up to date.
- Staff start each one by copying an old proposal.
- Quotes wait because only one or two people write them.
03 · A FIRST VERSION
What a first version looks like
An example of a focused scope. Your first version, price, and timing are agreed in writing for your project.
- One kind of quote or proposal, built from your template.
- A draft assembled from job notes, your current price list, and approved past examples.
- Prices and quantities taken from your list by ordinary rules, not invented by AI.
- Items the tool could not match, or assumptions it made, highlighted for review.
- A person reviews, edits, and approves every draft before it is sent.
04 · HOW WE’D DO IT
How we’d do it
The same four steps as our approach, applied to this kind of project.
Design
We agree on the template, sources, what AI writes and what comes from rules, the review step, the AI provider, and data handling.
Build
We build it and test it against past jobs before launch, comparing the drafts with the quotes that were actually sent.
Improve
We review real drafts with you, refine the guidance, and confirm the agreed criteria. Other document types are scoped separately.
05 · QUESTIONS
Common questions
Related: AI integration · Workflow automation
Could the AI quote the wrong price?
Prices come from your list through fixed rules, so the AI is not choosing figures. It can still misread notes, which is why unmatched items are highlighted and a person approves every draft.
Will it sound like us?
It works from your template and approved examples, so drafts follow your structure and wording. Expect to edit them, especially at first.
Where does our pricing information go?
Notes and past proposals are sent to the AI provider agreed in design, preferably under your own account. What is sent and kept is agreed before anything is connected.
How much does it cost?
It depends on the templates, the sources, and the tools it connects to, plus ongoing provider usage. Scope, price, and timing are agreed in writing after discovery.
LET’S BUILD SOMETHING USEFUL
Have a version of this problem?
Tell us how the work happens today and what a useful first version would change. A plain-language description is enough to start.