AI in Architecture and Engineering
Code checking, specification production and everything the practice already knows.
The AI conversation in architecture is dominated by image generation, which is the least useful application of it in a professional practice. The work that consumes fee and carries liability is specification production, code compliance and coordination, and none of it is glamorous enough to have attracted much attention.
Book a free consultationWhat makes this hard
- Code compliance checking is manual, jurisdiction-specific and where liability actually sits
- Specifications run to thousands of pages and are largely reassembled from previous projects by hand
- Detail and specification knowledge lives with senior staff and leaves when they do
- Coordination review across disciplines catches clashes late, when they are expensive
Where AI earns its place
Code Compliance Checking
Check drawing sets and specifications against the applicable code edition and jurisdictional amendments, flagging items for review rather than certifying anything.
Issues found before plan reviewSpecification Production
Assemble specifications from the practice's own master and prior projects, flagging sections that need active decisions rather than carrying forward defaults nobody chose.
Specs drafted from your mastersSubmittal and RFI Review
Check submittals against the specification section they answer, and draft responses to the routine portion of RFI traffic.
Days off review turnaroundPractice Knowledge Retrieval
Search details, specifications and decisions across completed projects in natural language, so the detail that worked in 2019 is findable.
Institutional memory that survives departuresFigures are drawn from Senteras engagements and are illustrative of typical results. Outcomes vary by data quality, infrastructure and scope.
The rules that shape the build
These are the constraints that decide the architecture, usually before anyone has picked a model. This is general information about the regulatory landscape, not legal advice on your obligations.
Professional licensure and seal requirements
A licensed professional takes responsibility for the work. AI assistance does not change who is liable, which means the workflow must keep a professional in the determination, not merely in the approval.
Building code jurisdictional amendments
Local amendments routinely override the model code. A system checking against the base code without amendments is confidently wrong in exactly the jurisdiction you are building in.
Standard of care
Professional liability turns on the standard of care, not on whether a tool made the error. Documented human review is what preserves the position.
Nothing here signs or seals anything
Every deliverable in this field is stamped by a licensed professional who carries personal liability for it. Senteras builds systems that assemble, check and flag, with a licensed professional making every determination. This is not caution for its own sake: a system that quietly makes a code determination creates exposure the practice cannot insure against and would not discover until a plan review or a claim.
Where this applies
The same core systems, with the differences that matter in each setting.
- Architecture practices
- Specification production and code checking; the payback is in fee protection on the least visible phase.
- Structural and MEP engineering
- Calculation checking and coordination review across discipline models.
- Civil engineering
- Permit documentation and jurisdictional requirement checking, which varies enormously by authority.
- Design-build contractors
- See the construction pillar; submittal and specification workflows converge.
Common questions
Will AI stamp or seal anything?
No. A licensed professional carries personal liability for the deliverable and makes every determination. Our systems assemble, check and flag. A system that quietly makes a code determination creates exposure the practice cannot insure against.
Can it check against local code amendments?
It must, or it is worse than useless. Local amendments routinely override the model code, so a system checking the base code is confidently wrong in exactly the jurisdiction you are building in. Amendment coverage is scoped per jurisdiction at the start.
What about all the AI image generation tools?
Useful for early concept exploration and largely irrelevant to where fee and liability sit. Specification production and code compliance consume the hours and carry the risk, and almost nobody is building for them.
How do we keep detail knowledge when people retire?
Retrieval across completed projects, so the detail that worked in 2019 is findable by someone who was not there. This is the use case practices consistently underestimate until a departure makes it concrete.
How we build it
Local & On-Prem LLM Deployment
The most powerful AI models, running entirely on your hardware.
Custom AI Agents & Automation
AI that doesn't just answer questions. It gets things done.
Model Fine-Tuning & Integration
Models that speak your industry's language, trained on your data.
Document Processing
Extraction, classification and validation across formats you do not control.
Internal Knowledge Base
Answers from your own documents, under your existing permissions.
Contract Review
Flags deviations from your playbook, not from a generic standard.
Start with a conversation, not a proposal
Thirty minutes. We will tell you what we would change first, and whether you need us at all.
Book a callThe firm behind the firm