Landscape Estimates and Renderings in One Client Record: An In-Person Build

A landscaping company was estimating in one long AI chat thread that kept losing context, pricing installed plants with the labor folded into the material price, and rebuilding a design template for every job. Then a client could not find the photos sent from the estimating software, and part of a deal was lost to disconnected delivery.

We ran an in-person build session with the owner, the sales estimator and designer, and the office manager. If you run one of the many home service businesses that quotes and renders for clients across several tools, this is the workflow we mapped.

The situation

The company's estimate and rendering work crossed five systems: GoHighLevel for the client record, job-costing software for the estimate, an AI assistant for quotes and renderings, a design tool for templates, and landscape design software for the renderings. Nothing tied them together into one client record.

What was broken

Four things kept costing time or money.

• Estimating ran in single long AI threads. Long threads lose context, so the assistant could lose earlier instructions and inputs.
• Pricing hid the labor. An installed shrub, for example, was priced with the labor folded into the material price, so there was no separate labor line.
• A design template was rebuilt for every job in the design tool.
• Delivery was disconnected. Photos sent from the estimating software were not found by the client, and part of a deal was lost because delivery was split across tools.

What we built

The goal was one client record that holds everything, so the next step is always visible.

• A two-way sync between GoHighLevel and the job-costing software, plus a pipeline stage called walkthrough completed that triggers the handoff.
• Three knowledge-base PDF files for the AI assistant: one for rendering logic, one for pricing logic and one for conversion messaging.
• AI projects with a separate quote chat and a separate rendering chat for each client, so no single thread carries every job.
• Work areas in the job-costing software with labor separated from materials and a 6% material markup, so the price shows what the labor and the materials each contribute.
• Combined client delivery: what the client needs goes out together from GoHighLevel, so the client gets it in one send.
• A design-tool plugin connected to the AI assistant, with one reusable empty template, so the template is built once.

What changed

This session was a build, and it did not report any quoting-time or close-rate figure. What changed is the shape of the workflow. A client now has one record, the quote and the rendering each have their own chat, and labor shows as its own line next to materials.

The combined send targets the delivery problem that cost part of a deal. Whether it prevents a repeat is something the next few quotes will show.

Check your own estimate workflow

This section is advice for your own business. Open your last quote and look for a line where labor is folded into a material price. If you cannot tell what the labor earned, split it out and add a markup you can see.

Then look at how your client receives the quote, the rendering and the photos. If they come as three separate messages from three tools, send them together from the system that holds the client record. Finally, give each client their own chat in your AI tool, so one long thread never has to remember every job.

Frequently asked questions

Why not keep every estimate in one long AI thread?

Long threads lose context, so the assistant can lose track of earlier prices, rules and inputs. A separate quote chat and rendering chat for each client keeps the working context small and tied to a single job.

How do you separate labor from material pricing?

Build work areas in your job-costing software with labor entered on its own line, then apply a stated material markup. Here that was a 6% markup on materials, so a plant is no longer priced with its labor hidden inside.

What belongs in an AI knowledge base for a landscaper?

Written logic the assistant can follow. We set up three PDF files: one for rendering logic, one for pricing logic and one for conversion messaging.

Why send the quote, rendering and photos together?

Clients can miss photos that arrive separately. Sending one combined message from the CRM keeps the record complete and puts everything in front of the client at once.

If your quotes, renderings and photos live in different tools, book a 30-minute AI follow-up and reporting audit. We map the path a client record takes and show you where a single record could hold it all.