Laura Dantonio

AI can generate synthetic users, actors and agents fast, but most teams still struggle to separate what is useful from what is risky, and how to defend these methods internally. In this hands-on workshop, Laura brings practical experimentation and Chiara adds a behavioural science lens to help teams evaluate, create and use synthetic representations responsibly.
You will compare approaches from synthetic personas grounded in first party research data, to scenario based synthetic actors and task oriented agents. You will pressure test outputs for bias, validity and misuse, and learn what “good” looks like in practice so you can make confident calls in your own context.

What you will learn

• The practical differences between synthetic personas, synthetic actors and task focused agents
• When each approach helps, and when it becomes misleading, across discovery, concept testing and higher risk domains
• How to structure messy research inputs to improve quality, traceability and repeatability
• Simple data science principles for building a scalable dataset for use with foundation models
• Prompt patterns, evaluation criteria and a lightweight decision framework you can use with cross functional partners

In this talk, I'm going to present a series of visual maps used to foster cross-discipline collaboration and team alignment in an enterprise search redesign. Search is a complex experience that requires a highly technical team and cross-discipline collaboration. For this project, I created a search framework to understand users' pain points in a visual map to guide design/development and foster collaboration. We used this framework to redesign search experiences and found it useful to prioritise search features and support collaboration. Participants will learn why using visual maps is useful for team collaboration and they will be inspired to create maps with practical examples.