Sealed Opinions, Code & Fees: A 2026 Architecture Firm Playbook

Design and engineering are consultative by nature, which makes a website assistant feel like a natural fit: answer questions, qualify projects, book consultations. The danger is that the questions clients ask first, is this wall load-bearing, will this meet code, can we build this, are not casual. They are professional opinions, the kind licensing laws reserve for a licensed architect or engineer who has examined the actual project, because people build and live based on the answers. An assistant that obliges is performing licensed work on the firm’s behalf.

This guide is the companion to the threat side of that story. The threat piece covers what goes wrong when an unguarded chatbot renders structural and code opinions. This one is the standard: what a compliant deployment looks like for an architecture or engineering firm in 2026, and the specific lines the system has to hold.

Structural and Code Calls Are Licensed Work

Determining whether a wall is load-bearing, whether a design satisfies the building code, or whether a structure is sound is the practice of architecture or engineering, and state licensing laws reserve it for licensed professionals. The seal on a drawing exists for exactly this reason: it marks an opinion a qualified, accountable person stands behind. An assistant that tells a homeowner a wall is "probably fine to remove" or that a plan "meets code" is performing sealed work without a license and without seeing the building, on a client who may act on it with a sledgehammer. The harm is concrete, a dangerous renovation, a failed inspection, a structure that has to be torn out, with the firm’s website on record as the source.

Feasibility Answers Are Opinions Too

The softer-sounding questions carry the same weight. Will this addition be allowed, can we build to the lot line, will the city approve this, these depend on zoning, local code, and site specifics a website cannot evaluate. An assistant that says "that should be approvable" is making a feasibility judgment the client will rely on to spend money, and the gap between that optimism and the jurisdiction’s actual rules is the firm’s problem. Feasibility is a professional opinion, not a quick reassurance.

Fees and Scope Are Commitments

The business side has its own exposure. "A project like that runs about this much" and "we can have drawings to you in three weeks" read as commitments, and a client who plans around them has relied on terms the firm never set. A promise the firm cannot keep can also draw scrutiny under the FTC’s authority over unfair and deceptive practices. Fees, timelines, and scope belong in a proposal a person prepares.

The 2026 Compliance Standard, Line by Line

A compliant architecture and engineering assistant is defined by what it is built to refuse. Treat the list below as the floor.

• No structural or code opinions. "Load-bearing" and "meets code" are licensed judgments, so the assistant books a consultation instead.
• No feasibility verdicts. "That’ll get approved" depends on zoning and site specifics a person evaluates.
• No fee or timeline commitments. Those belong in a proposal, not a chat window.
• No standing in for the seal. The assistant never renders the professional opinion a stamp represents.
• Every exchange is logged, so what a client was told is reviewable rather than lost in a widget.

The pattern is the same one that runs through every vertical. The assistant answers what carries no obligation, services offered, the firm’s process, project types, how engagements work, general "how does a project like this go" questions, and routes structural calls, code opinions, feasibility, and fees to a licensed professional.

Why an Instruction Cannot Meet the Standard

The usual shortcut is to write the rules into the assistant’s prompt. Tell it never to give professional opinions, never to quote a fee, and call the boundary set.

It is not set, because of how the model reads a request. It follows an instruction when the question resembles the wording it was warned about, and slips the moment the phrasing changes. You tell it never to give professional opinions. The client never says "give me a professional opinion." They ask, "is the wall between these rooms structural, and would an open plan pass?" The model hears a friendly question and answers it. The instruction was loaded the whole time. It just did not recognize the sentence that became a structural and code opinion.

That is the difference between an instruction and a standard. An instruction asks the model to behave; it does not stop the model from speaking. A real boundary is built into the system and decides what the assistant may say before it answers, so a structural call, a code opinion, or a fee never reaches the client no matter how the question is framed. "Will not" is a suggestion. "Cannot" is an architecture.

What a Compliant Deployment Looks Like

Meeting the 2026 standard does not mean a brochure site with no assistant. It means deploying one built to qualify projects and book consultations without rendering professional opinions, and one that keeps a clean record of what clients were told.

Fred is built that way. It answers from your own firm content, captures the project, and routes structural calls, code opinions, feasibility, and fees to a licensed professional. It runs more than 50 industry guardrail packs, and the architecture and engineering pack is built around the practice-of-licensure line, the seal, and fee commitments. Fred does not say a wall is safe to remove, that a design meets code, or what the project will cost. It cannot. It answers what it should, captures the project, and hands the professional judgments to the people who can stand behind them.

That is the difference between hoping the assistant does not overstep and being able to show why it cannot.

Frequently asked questions