Working with AI: How to use AI without losing your brain

Fendi Tsim

Behavioural Scientist, SCAN co-creator

Alina Gutoreva

Assistant Professor of Cognitive Science

+ Jason Vu Nguyen and Sofia Pires

Apply SCAN to your work, identify when AI builds value and when it drains it.

Knowledge workers rely on one thing above all: their thinking. Team leaders and clients pay for judgement that comes from experience and expertise. AI is now quietly reshaping our skills and if used indiscriminately, it can erode what differentiates us from machines and other humans.

The pressure to use AI more is everywhere. What's missing is a clear way to decide, task by task, when it helps and when it costs your long term thinking and your place at the table.

This workshop gives you SCAN, a decision-making framework designed to help knowledge workers determine how to assign tasks when working with AI.

You'll apply it to your own current work first, sorting real tasks into where AI should lead, where it should assist, and where it has no place at all.

Then you'll turn the same framework outward, building it into something you can offer clients directly, whether that's a diagnostic, a short session, or a piece of your existing advisory work.

You'll leave with sharper judgement about your own AI use, and a packaged way to bring that judgement into your client relationships.

What you’ll learn

Learn SCAN, protect the thinking that makes you valuable, and turn it into a service worth paying for.

Walk through the four SCAN zones, Substitute, Complement, Aid, Non-negotiable, with real examples.
A simple set of diagnostic questions you can run through in seconds on any new task.
Practice sorting a mix of sample tasks live, before applying it to your own.

Live worksheet where you map three current tasks against the SCAN framework.
Peer discussion and knowledge sharing.
A template for structuring SCAN as a client-facing session or diagnostic.

Digested research behind cognitive offloading and automation bias, explained plainly and without jargon.
Case study of AI use tipping from useful into corrosive, so you recognise the pattern rather than just the theory.
A simple way to track which tasks are drifting toward over-reliance, so you catch it early.

Workshop agenda

The SCAN framework (30 min)
- Walk through the four zones: Substitute, Complement, Aid, Non-negotiable. - Real examples for each zone, not abstract theory. - Diagnostic questions you can run in seconds on any task.
Applying SCAN to your own work (90 min)
- Live worksheet, map three of your current tasks against the framework. - Surface the tasks that are harder to classify. - Peer discusion, compare where people land and why.
Packaging SCAN for clients (60 min)
- Turn the framework into a client offer: diagnostic, short session, or add-on to existing advisory work.