Best AI for architects: a practitioner's map of AI for architecture

The render gets the client to yes. The forty emails after it are where the project is actually won or lost. Most of what is sold as AI for architects is about the first part. Below: the three kinds of tool you will meet, with named examples and the prices their own pages publish, what the AIA's AI Task Force has said about using them, and where the client side of a project fits.

AI in an architecture practice does three separate jobs, and most guides only cover the first two. There are rendering and concept tools that turn a sketch or a model view into images. There are assistants built into the design and BIM software you already use. And there are general assistants for writing, summarising and thinking out loud. What sits outside all three is the client side of a project: the budget worry the client mentioned at the last review, the contractor's question that arrived while you were on site, the revised drawings you promised by Friday.

One disclosure before going further. This guide is published by Equerry, a personal assistant for that client side. Equerry does not design, draw or render anything. It does not open your model, produce images or check drawings. The first part of the page covers the landscape on its own terms; Equerry comes in near the end, where it fits.

The three kinds of AI an architect runs into

Named tools by kind, with the prices their own pages publish

The table below names current examples of each kind, described in each vendor's own words and with the price its own page showed on 2026-10-27. It is a map, not a ranking: the tools are listed by kind, there are others in every kind, and none of them is a recommendation. Prices exclude tax, change often, and in some cases depend on the region you browse from; where a vendor's page disagrees with this table, the vendor's page is right.

Rendering and concept tools

This is the category most searches for AI tools for architects are really after, and it is the most crowded. The four named above differ mainly in which design software they plug into (Veras lists seven hosts; ArkoAI and mnml.ai name Rhino, Revit and SketchUp; Rendair starts from a sketch, plan or model without naming a host) and in how usage is metered, which is almost always credits per image. Chaos makes the client case for the whole category on the Veras page: "Walk into every client meeting with multiple options." Vendors publish prices in different shapes, so compare the cost of a month of your actual output rather than the headline figure.

The useful discipline is the same whichever you pick: treat the output as a sketch, not a design. An image that looks finished can suggest decisions nobody has made yet, so say so when you show it to a client.

Assistants inside design and BIM software

The major design platforms are adding assistants of their own, and the two named above show the pattern: the assistant arrives inside software the firm already licenses, is included with certain plans rather than sold on its own, and is often still labelled a test version while it settles. Graphisoft's is a documentation-and-model assistant; Trimble's SketchUp AI bundles rendering, geometry generation and help. Because these sit inside your existing licence, the practical questions are usually about availability on your plan and what happens to project data, and the answers are in each vendor's own documentation.

General assistants

ChatGPT, Claude and Gemini are where many architects start, because they are good at words: a first draft of a fee proposal, a plain-language explanation of why the stair moved, a summary of a long consultant report. They are not design tools. What each one knows about a project depends on what you type, upload or connect to it, and on its settings. Connecting a whole mailbox or drive gives a general assistant more to work from, and it also widens what it can reach, which is a decision to take with your firm, not on your own.

By tool: ChatGPT for architects covers what architects actually use ChatGPT for, what the AIA Task Force guidance says as published, and what never to paste.

What the AIA's AI Task Force has said

The American Institute of Architects' AI Task Force, in an article published April 6, 2026 (re-read 2026-10-27), set out how it sees the tools. The article is short and worth reading in full. Three points bear on how you use any of the tools above:

• What the tools are good at today. "Currently AI is effective at augmenting creative capabilities and handling repetitive, time-consuming tasks that take up professional time without drawing on professional judgment."
• Where the time goes instead. "The architect's focus can shift to those areas that require more human expertise: contextual thinking, creative problem-solving, and the client and professional relationships that make good architecture possible."
• Oversight stays with people. "Human oversight and professional accountability are the critical factors that make AI useful and less risky."

The same article refers to the AIA Position Statement and Responsible Use Guidance published in January 2026, and points readers to the AI Task Force's resource centre. That material is AIA's view for its members; this page makes no claim about what the AIA permits or requires. Your firm's own policy, your contracts and your clients' expectations are what govern your work. This page summarises published information and is not professional advice.

The part nobody covers: the client side of a project

A residential extension or a small commercial fit-out runs on correspondence. The client emails about the kitchen layout on a Sunday night. The contractor asks about a window detail. The structural engineer's comments arrive as a PDF. The planning consultant replies to a thread you were copied on three weeks ago. Somewhere in that flow is the one sentence that decides the next review: "we can live without the rooflight if it keeps us under budget."

You can keep that record well without any AI, and good practices already do some version of this:

• Write the meeting note while you are still in the car. What was decided, what the client asked for, what you owe and by when, and the one remark you will want to quote accurately later. Our guide to organizing meeting notes covers a format that holds up.
• Send the recap email the same day. A short summary after a client review puts the decision where everyone can find it and invites a correction while memories are fresh.
• Photograph the site, then caption it. A site photo with no note is a puzzle three months later. One line on what it shows and why you took it is enough.
• Brief yourself before each review. Ten minutes with the last meeting note, the latest emails from the client and contractor, and the open items.

The weak point is retrieval. The note exists, the photo exists, the email exists, but they live in three places, and the reply that changed the position arrived after you wrote the note.

Where Equerry fits: a personal assistant for the client side

Equerry is a personal assistant who helps you get things done, and for an architect that means the client correspondence around the design, not the design itself. It keeps track of what you share with it and briefs you from it, the way a good project assistant keeps the file straight. It does not read drawings or BIM models and connects to no design software.

• CC it on the project threads. Equerry has an address of its own at mail.getequerry.app. Forward it a thread with a client or contractor, or add it to the CC line of a live one; forwarding delivers the thread as it stands, and keeping Equerry on the CC line is what brings later replies in as they arrive. It never asks for access to your inbox, and the original stays in your own mailbox. What Equerry remembered from it is listed in Settings → Your memory, each line marked as from that email, and you can delete any line.
• Send it site photos and PDFs. From your iPhone's share sheet you can send a site photo, a screenshot of a marked-up drawing, a PDF of the engineer's comments or a link. You can also type a note in the app.
• Debrief the review by voice. Walking out of a client meeting, record a short voice note in Equerry: what was agreed, what the client pushed back on, what you owe. It transcribes it and remembers it. Equerry does not join or record your meetings; the voice note is yours, made afterwards.
• Ask in plain language. "What did the client say about the rooflight?" is answered from the emails, voice notes and notes you gave it. When an answer draws on a note you saved, it carries a source chip: tap it to see the passage Equerry used and open the original.
• Walk into the next review briefed. If you allow calendar access, Equerry keeps a snapshot of your events from today through the next 14 days (titles, start and end times, locations and attendee display names, never event notes, meeting links or email addresses). If you set up Meeting Prep, before a client review it briefs you from what you have shared: what the client last wrote, what is still unanswered, and what you owe them from the previous meeting.

Some honest limits. Equerry only knows what you send it, plus your calendar if you allow it; it cannot find the email you never forwarded. It does not open drawings or models and does not connect to design, BIM, project-management or accounting software. Anyone on a thread can see the Equerry address on the CC line, so decide per thread whether it belongs there. And sending something is still sharing it: what you send is processed to make it searchable and answerable, including by the outside providers named on our AI disclosure page. Whether client material goes into any tool, this one included, is a call for you under your firm's policy and your client agreements. What Equerry changes is scope: the question is about the specific threads, photos and notes you chose to send, rather than a whole mailbox.

If most of your client work happens in email, read our guide to choosing an AI email assistant next. For ChatGPT specifically, see ChatGPT for architects. For the broader picture of client-facing work, see Equerry for client-facing executives.

Sources, all read 2026-10-27 unless noted. The American Institute of Architects, AI Task Force, Architects and AI: Practical guidance for a changing profession (April 6, 2026). Chaos, Veras. Rendair, home page and pricing. mnml.ai, home page and pricing. ArkoAI, home page and pricing. Graphisoft, AI Solutions. Trimble, SketchUp AI and SketchUp plans and pricing; Trimble news, Trimble Launches SketchUp AI (December 10, 2025). OpenAI, What is ChatGPT Plus? (help centre, read 2026-10-27). Apple, ChatGPT on the US App Store (in-app purchases). Anthropic, Claude pricing. Google, Google AI plans (US). Vendors change their products and pages often; where a vendor's current page disagrees with this one, the vendor's page is right. This guide is general information, not professional advice.

Last updated 2026-10-28.

All product names, logos, and brands are property of their respective owners. Equerry is not affiliated with, endorsed by, or sponsored by any other product mentioned; comparisons are provided for informational purposes.