ChatGPT for Architecture: Project Briefs, Research, and Review Preparation

ChatGPT can help architects write project briefs, research building types and precedents, draft client communications, and prepare for design reviews. It does not design buildings, produce drawings, or replace the professional judgment required for code compliance and life safety. Tools like fenbs carry the approved requirements and design decisions between sessions so each new conversation picks up from the current state of the project rather than starting over.

Architecture projects run for months or years. The challenge is not only producing good work but keeping complex information accessible — to the project team, to consultants, and to AI assistants that work alongside them. That shared record is what makes AI assistance sustainable across a long project.

Developing project briefs

A project brief translates a client's goals into a set of design requirements: space needs, functional relationships, technical requirements, budget parameters, and schedule constraints. Drafting it from client conversations and intake notes is time-consuming. ChatGPT can structure that material into a brief once you supply the key facts.

Paste in your meeting notes and a list of the known constraints, then ask for a draft brief organized by functional area, technical requirement, and open questions. The draft surfaces gaps: a missing adjacency requirement, a budget constraint that was implied but not confirmed, or a program element the client mentioned once and has not discussed since.

Send the draft back to the client to confirm before design work begins. A confirmed brief is the anchor that holds the project together when scope discussions arise later.

Researching building types and precedents

Early-stage research on a building type — relevant design precedents, typical space relationships, common technical challenges, applicable standards — helps frame a project before design begins. ChatGPT can summarize publicly available information on a building type and help you structure a research outline for the project.

For code and standards research, use the primary source. International Building Code, local amendments, ADA standards from the US Access Board↗ (opens in a new tab), ASHRAE standards, and state-specific requirements change and have nuances that a language model may not reflect accurately. ChatGPT can explain the general framework; a code consultant or the authority having jurisdiction is the authority for your specific project.

For historic preservation projects, local landmark regulations and Secretary of the Interior's Standards for Treatment of Historic Properties govern what is permitted. Check those requirements directly rather than relying on an AI summary. For related AI work in the design field, see AI design assistants for project work.

Preparing for design reviews

Design reviews — internal, client, planning board, or permitting — require clear communication of design intent, responses to previous feedback, and explanation of how the design meets the stated requirements. ChatGPT can help structure a presentation narrative, draft a written design description, or prepare a response to previous comments.

For a presentation, describe the design decisions and their rationale; ask ChatGPT to draft an explanation for a non-technical audience. For a response to previous comments, paste in the comments list and a summary of how the design addresses each one; ask for a structured response document.

Planning board and permitting submissions have specific format requirements set by the jurisdiction. The substance of those documents — the drawings, the required forms, the technical calculations — is professional work. AI can help with narrative sections and cover letters, not with the technical and legal content of a formal submission.

Client and consultant communications

Architecture projects involve ongoing communication with clients, structural and MEP engineers, contractors, and authorities. ChatGPT can help draft meeting summaries, scope clarification letters, and RFI responses from your notes. Review for accuracy before anything goes out — a scope letter that misrepresents what was agreed is a contract problem.

For construction administration, field reports and RFI responses carry professional and sometimes legal weight. These still need the architect's review and judgment. AI-assisted drafting compresses the writing time, not the responsibility.

Professional limits and the architect's responsibility

The architect of record is professionally responsible for the design under state licensing laws. AI does not hold a license. Structural calculations, life safety analysis, accessibility compliance, and energy code compliance require licensed professionals and verified calculations. ChatGPT can help you think through a problem; the responsible professional makes the determination.

For the AIA's guidance on professional practice, see AIA's practice resources↗ (opens in a new tab). For project management in an AI-assisted workflow, how to use AI at work covers general principles.

When the client approves a design direction, a material specification, or a program change, record it in fenbs Decisions and rules. Every connected assistant reads those decisions before acting, so a session a month later starts from the approved state rather than the previous proposal.

How to do this with fenbs

1. Create a free account. Add a project for each active commission.
2. Add tasks for each phase: programming, schematic design, design development, construction documents, permitting, construction administration. Move tasks through the board as phases progress.
3. Open Decisions and rules. Record the confirmed program requirements, approved design directions, and key decisions the client has made. An AI assistant reads these before acting.
4. Connect your AI assistant from Integrations. At the start of each session, ask it to read the project context.
5. After each client meeting, update Where we left off: what was agreed, what was deferred, and what the next step is. The next session picks up from there.
6. Use Lessons learned after project completion: what programming and scoping revealed, what made the review cycle efficient, what would change the next time.