How Landscape Architects Are Using AI for Client Documents in 2026

How landscape architects are using AI for client documents shows up clearly in the survey data: the writing comes first, and the drawings stay with the designer. Proposals, briefs, reports and RFP responses are where studios hand AI a first draft, while planting plans and grading stay in the tools built for them.

TL;DR: Landscape architects use AI mostly for the written layer of a project: briefs, reports, proposals and RFP responses, plus reading long client briefs. Surveys by ASLA, CSLA and IFLA put drafting of briefs and proposals at 47% of respondents, and small firms report that AI produces a rough starting draft they rework heavily. The step that remains is layout. DocsAura, an AI document design tool, takes the finished DOCX or PDF and returns a designed page in about two minutes.

How Landscape Architects Are Using AI for Client Documents Today

Three sets of numbers describe the profession's current habits.

The ASLA Digital Technology Professional Practice Network ran an AI survey in 2024 and received over 300 responses. It found 55% of respondents using AI in practice, teaching or research. Among design-process uses, writing reports came to 45%, conceptual design 41% and responding to RFPs 29%.[^1]

The 2025 AI in Landscape Architecture survey, run by an academic team with CSLA, ASLA and IFLA, points the same way. Its highlights list background research and information gathering at 50%, drafting briefs, proposals or syllabi at 47%, and predesign or business development work at 41%. Three-quarters of respondents (75%) reported increased task efficiency, and 92% learned the tools through trial and error.[^2]

A third reading comes from Landscape Architecture Magazine, which in April 2026 quoted a researcher at Utah State University describing a shift from an early period of experimentation with generative AI toward using AI in practice, particularly in text-based workflows.[^3]

Put together, the pattern for a small studio is specific. Chat assistants write the first pass of documents the studio already produces. Design tools make images. The two rarely meet, and the client receives whatever the first pass becomes.

Where Client Documents Sit in a Landscape Project

A landscape project produces a paper trail at every stage. Damian Holmes, founder of World Landscape Architecture, lays out nine stages in his practice guide, and four of them end in a document a client reads:[^4]

1. Return brief or proposal. After the first consultation, the landscape architect sends a document covering the project understanding, scope and deliverables, team, consultants, fee and programme. Holmes notes it can run from one or two pages to 100.
2. Concept design report. Plans, site analysis, precedent images, renders and material palettes that show the client the proposed arrangement, functions and overall look.
3. Design development report. The concept developed in more detail, with the issues found so far (topography, access, services, drainage, budget).
4. Maintenance and defects reports. During the maintenance period the landscape architect is often required to attend and submit reports.

Each of these starts as text the studio writes from its own notes. That is the layer where the surveys show AI use, and it matches the studio's weekly reality: the drawing set is the product, and these documents are the packaging that goes with it.

What a 14-Person Firm Reports

Landscape Architecture Magazine spoke with John Payne, ASLA, a partner at SiteWorks, a 14-employee firm. His staff put an RFP into ChatGPT and use carefully built prompts. His estimate of the result is plain: the output "likely generates about 20 percent of the needed document," and the team still alters about 80 percent of the text, fleshes out ideas and personalizes it. On one RFP for an operations and maintenance planning project, he put the saving at about an hour, and he said skipping that busywork left his creative energy for the design thinking.[^3]

That is a useful benchmark for an owner deciding whether to try this. An hour per RFP, with the thinking still done by a person, adds up across a year of bids. It also matches what ASLA's own survey found about time: 27% said AI had saved them time, 48% were unsure and 7% said it added time.[^1] Studios that track the saving can see where it lands. Studios that don't track it cannot say.

Reading the Client's Documents Is the Other Half

Some of the clearest gains come from the inbound side. Henning Larsen, a 600-person firm with 120 landscape architects, uploads the client's brief and project information into NotebookLM or Perplexity so the team can ask questions of it instead of searching for keywords. Its innovation and sustainability director gave an example: a bicycle lane width buried deep in an RFP can be found by asking "How wide should the bike lane be?" The firm reports that accuracy is generally high from a closed set like a brief, that the tool cites its sources, and that a person reviews the answers.[^3]

On a campus project for the Max Planck Institute for Demographic Research in Rostock, the client sent 258 pages of German text. On a Singapore project, a brief that would have taken two people two days to read and translate for the team took one person an hour and a half.[^3] A 600-person firm has an innovation team behind that workflow. A five-person studio can borrow the habit with the same rule: AI reads, a person checks the answer against the page.

What we found reading four landscape proposal guides

On 4 November 2026 we read four of the guides that rank for landscape proposal templates: a fee-proposal guide from a software vendor, a landscape proposal guide from a contractor-software company, a proposal template from another software company, and a blog tutorial on writing a landscape design proposal with AI. Three of the four give exclusions their own section ("not included"), and three of the four price the work by named stage or phase. Two of the four mention AI at all, in both cases for drafting the wording or customizing a template. None describes starting from a proposal the studio has already written and turning it into a designed page, and none covers the layout step for the written documents that go with a drawing set.

The Step After the Draft: Turning the Text Into a Page

The studio's writing is done. The fee proposal reads well, the scope has its stages and exclusions, and the concept report has its narrative. What remains is a Word file that looks like a Word file, sitting next to a drawing set that looks like the work of a design practice.

This is where DocsAura, an AI document design tool, fits. It takes a DOCX or PDF the studio has already written and returns a designed HTML page in about two minutes. It does not draw, model or produce plans, and it has no link to AutoCAD, Rhino, SketchUp or Revit. Plans, renders and planting schedules come from the tools the studio already uses, and DocsAura handles the pages of text around them.

The pitch is time. A landscape architect knows how to make a page look right, and the evening goes on doing it for the third document of the week. Handing that evening back to the project is the point of the tool.

Two Things to Settle Before the Documents Go Out

Two findings from the ASLA survey deserve a decision from every studio owner.

Disclosure. Of the respondents, 56% said they do not include disclosure when AI is used in deliverables, and only 7% said their firm permits AI-generated content in stamped plans.[^1] Clients are also asking. A digital studio manager at Confluence told Landscape Architecture Magazine that clients now expect firms to say whether and how they use AI, even unprompted, and that firms which stay silent get blowback.[^3] A one-line statement in your proposal terms settles it for most clients.

Confidentiality. A proposal for a private client contains a name, an address and a budget. Our guide on whether it is safe to upload architecture client documents to AI sorts studio paperwork into green, amber and red by content, and the same sort works for a landscape studio. For the mechanics, see what to redact before uploading a document to AI.

A Routine for the Next Proposal

1. Draft the text yourself or with a chat assistant. Give it your stages, deliverables and exclusions, and expect a starting point. SiteWorks puts it at a fifth of the finished document.
2. Rewrite until it sounds like the studio. The estimate above says 80% of the text changes, and that is normal.
3. Check every number and name. Fees, areas, dates and plant names come from your records.
4. Sort the document green, amber or red before it goes into any tool.
5. Design the page. Drop the finished file into DocsAura and read what comes back.
6. Send it with the AI line in your terms if your clients expect one.

If you are writing the fee proposal itself, our walkthrough on how to use AI to write an architecture fee proposal covers the stages and exclusions in detail, and the stage names carry over to landscape work.

Try It on One Document You Already Have

Open the last proposal or concept report you sent, the one you wrote in an evening. Drop that one file into DocsAura, the AI document design tool, and see what comes back in about two minutes. There is nothing to install and no template to build. If the page suits the studio, send it next time. If a section needs a change, edit it in the browser.

[^1]: ASLA, "How Landscape Architects Are Incorporating Artificial Intelligence: Insights from the 2024 ASLA Digital Technology PPN Survey," asla.org, read 4 November 2026. [^2]: International Federation of Landscape Architects, "2025 AI in Landscape Architecture Survey," iflaworld.com, read 4 November 2026. [^3]: David Tobenkin, "Your Mileage May Vary" (web headline: "'We Jumped in With Both Feet': Designers on AI's Back-Office Impact"), Landscape Architecture Magazine, 22 April 2026, landscapearchitecturemagazine.org. [^4]: Damian Holmes, "Practice | Stages of a Landscape Architecture Design Project," World Landscape Architecture, worldlandscapearchitect.com, read 4 November 2026.