AI Sales Quoting Assistant for Landscapers: Internal Review vs Client-Ready Quotes

One AI quoting tool was being asked to do two jobs that pull in different directions. The estimator needs a blunt internal view of cost and margin. The customer needs a clean, client-ready proposal. At one landscaping company the same tool had to serve both, and its formatting, its pricing references and its way of capturing field measurements were not yet reliable.

We ran a sales GPT training call with the owner of a landscaping company, an estimator and designer, and a crew lead. If you run one of the many home service businesses that quote on site, the structure we settled on is a useful template.

The situation

The company's quoting assistant is a custom GPT with a knowledge base built from the company's own pricing. The training call had one main aim: make the system faster to use in the field while keeping quote structure, pricing logic and client-facing presentation consistent.

That sounds like a formatting job. It is really a question of audience. A quote the estimator reads to check margins should show costs, assumptions and flags. A quote the customer reads should show scope and price, not the internal numbers.

What was not reliable yet

Three gaps showed up on the call.

• Formatting: quotes needed one consistent structure.
• Pricing references: the tool needed trustworthy sources for material, plant and fertilizer pricing, and those sources were still being gathered.
• Field inputs: the team needed a more reliable way to capture measurements and quote inputs on site.

What we set up

Most of the work was about what the tool reads and how it answers.

• Two outputs: internal review output is separated from client-ready output, so margin detail does not have to be edited out of a customer proposal.
• A standard quote structure: quote structure, pricing logic and client-facing presentation were standardized, along with process updates for how quotes are organized.
• Real pricing references: pricing documents, material lists, plant pricing and a photographed fertilizer pricing sheet were gathered to update the knowledge base and the formatting rules.
• A reviewed baseline: the owner and a second team member were to go through the quoting-guide draft together and confirm the latest pricing before it became the new baseline for the GPT.
• A follow-up review: a session later that week with the team to review the updates and the next round of refinements.

What changed

No quote-accuracy or time-saved result was reported. The sources were still being gathered when the call ended, and the review was set for later that week.

What changed is the design of the tool. The plan from the call is one assistant with two outputs: an internal one for checking margin and a client-ready one for sending. The pricing it relies on was being pulled into a reviewed set of documents with two named reviewers.

What to take from this

This section is advice for your own quoting setup. Decide who reads each output before you write a single prompt. Put the internal numbers in one view and the customer-facing scope in another, and never ask the same screen to do both.

Then treat your pricing documents as the product. A photo of a printed price sheet is a fine source if it is clear and current. Pick someone to review the baseline before the tool uses it, and decide which measurements your crew must capture on every job.

Frequently asked questions

Why separate internal and client-ready quote output?

The estimator needs costs, margins and assumptions, and the customer needs scope and price. Here the internal review output was separated from the client-ready output so the same tool could serve both without margin detail appearing in a proposal.

What should go into a quoting GPT knowledge base?

Your real pricing references. In this case that meant pricing documents, material lists, plant pricing and a photographed fertilizer pricing sheet. Keep them current, and have someone confirm the latest numbers before they become the baseline.

How do you keep field measurements reliable in an AI quote?

The call aimed at a more reliable workflow for measurements and quote inputs. Our advice is to decide which fields the estimator must capture on every job before the tool is allowed to quote, so a missing measurement is caught at the start.

Who should review the pricing baseline?

Here the owner and a second team member were to go through the quoting-guide draft together and confirm the latest pricing before it was loaded. Two people who know the prices catch stale numbers that one person would miss.

If your quotes are built differently by each person or shown to customers with internal numbers still in them, book a 30-minute AI follow-up and reporting audit. We review how you quote today and show you what a two-output quoting tool would look like for your business.