Practice area one
AI Trust, Safety and Governance
Where AI decision support belongs, how much to rely on it, and how to govern it, for emergency services, healthcare and other organizations where decisions under pressure are the job.
Why it matters
Appropriate reliance, not maximum trust
Organizations adopting AI decision support need people to rely on it when it is right and question it when it is not. Research on trust in automation describes two ways this goes wrong: misuse, where people lean on a tool past the point it deserves, and disuse, where they set aside a tool that would have helped (Parasuraman & Riley, 1997; Lee & See, 2004). Automation also changes the human's job, often in ways that make the hardest moments harder (Bainbridge, 1983). Getting this right is a human factors problem before it is a technology problem.
Services
What we do
Pre-deployment assessment
Before a tool goes live: the decisions it will touch, how it can fail, what the people using it need to know, and a Do Not Touch list of where AI should not go.
Trust and appropriate reliance
How much your people rely on a tool, and whether they rely on it for the right reasons. Workflow and interface changes that make good judgment easier.
Reliability and safety assessment
Testing a tool against the conditions it will meet in real operations: the busy night, the unusual call, the degraded system, not only the vendor's benchmark.
Governance frameworks
Policies, accountability, procurement criteria and oversight that hold up on a busy shift, not only in a report to council or the board.
In-service monitoring
Tracking overrides, errors and drift once a tool is running, so problems surface early and lessons feed back into design and training.
Implementation support
Piloting with real users in real workflows before scaling, with training and change management built in from the start.
Start here
The AI Decision Architecture Workshop
One day, on site, with your leadership team in the room. In six structured sessions we map how decisions actually get made in your organization, identify where AI and automation fit well and where they do not, and produce a Do Not Touch list specific to your operations. You receive the deliverables and a 90-day roadmap within five business days. The workshop stands on its own, and for organizations ready to go further it becomes the scoping document for a full engagement.
AI Decision Architecture Workshop
Ask about the workshop
Tell us roughly how many leaders would attend and what you most want to look at. We will come back with an outline and the fee.
Our basis
Grounded in research
This practice area draws on Scott Ramey's doctoral research at the University of Waterloo on trust calibration and appropriate reliance in AI-assisted decision support. Our work uses published science, never unpublished study data. The Human Factor is independent of the University of Waterloo.
Not sure where to start?
The AI Readiness Assessment takes about five minutes and gives you a personalized report. Or tell us what you are weighing and we will set up a 30-minute call.
Book a 30-minute call
Talk to us about AI
What tool are you weighing, and who will rely on it? Thirty minutes, no charge.
References
Bainbridge, L. (1983). Ironies of automation. Automatica, 19(6), 775–779.
Lee, J. D., & See, K. A. (2004). Trust in automation: Designing for appropriate reliance. Human Factors, 46(1), 50–80.
Parasuraman, R., & Riley, V. (1997). Humans and automation: Use, misuse, disuse, abuse. Human Factors, 39(2), 230–253.