If our engineering team uses AI to verify structural calculations and the tool misses a critical error that goes to a client, how do we handle the liability and operational fallout within our leadership team?

Question

Tyler Smith · Accepted Answer

When an AI tool makes a mistake that slips through to a client, you cannot blame the machine, and you cannot blame the developer who built the tool. Under the EOS framework, accountability always stops with the human seat holder on your Accountability Chart.

If your engineering team missed a calculation error because they trusted the AI blindly, the Issue must be brought to your weekly Level 10 Meeting for a thorough IDS session. The first step is to identify the root cause. Did the employee skip the verification step outlined in your documented core processes, or was the process itself poorly defined?

If the employee skipped the required check, you have a People Component issue. You must address whether they still get, want, and have the capacity to do the job. If the process itself was at fault, you must immediately update your documented core processes to build in a stronger human audit mechanism. You must make sure that no automated calculation is ever accepted as final without a qualified human engineer signing off on the raw numbers. Treat the software like a junior apprentice: valuable for drafting, but never trusted to work unsupervised. This keeps your quality standards high and your team fully accountable.