At the 2025 SEAOC Convention in San Diego in September, members of NCSEA’s AI Grant Team presented “Advancing AI in Structural Engineering.” The session highlighted practical ways structural engineers can securely integrate AI into their workflows—emphasizing that AI is a tool to support, not replace, engineers. The adoption path was simple and deliberate: adopt an AI policy, get secure AI into the hands of your engineers, run small pilots, then expand in modular increments so tools map to actual workflows. The policy matters because confidentiality, risk, and misuse need to be addressed at the start.
The grant team approach to working with AI starts with data. When project knowledge is stored in managed cloud systems and organized into searchable knowledge bases, AI becomes useful in document‑heavy and early design tasks. Examples we showed included plan checks that flag sheet and schedule mismatches earlier, automation that assembles submittals and field reports from standard templates, quick generative studies to explore structural options, and RFI triage that classifies, routes, and drafts context‑aware responses for engineer review.
The team also covered how teams can build lightweight internal chat apps that sit inside existing tools and draw from project files with access controls. Several low‑code approaches make it realistic for small groups to prototype safely. On the coding side, AI is now a practical accelerator. Code assistants can generate boilerplate, help write small connectors between systems, refactor scripts, and explain errors so engineers can create or maintain utilities with faster turnaround. Engineering modules can be created faster than ever and orchestrated together in automation routines. Agent workflows have also matured in 2025, which means goal‑driven sequences can reliably handle narrow, multi‑step tasks like processing RFIs or standardizing report packages, with the engineer staying in the loop.
Platform choices should be evaluated on security, retention, and seat requirements, with attention to avoiding overlap. Guardrails remain essential. The engineer in Responsible Charge stays engaged from start to finish, and after‑the‑fact review alone is not sufficient. The expected gains are incremental, but they are achievable now where data and workflow fit are strong. The time is now for firms to engage with these tools, responsibly, and determine how they can help add value to your firm and the profession.