2026 Agenda

 

Export | Need Help?
All Sessions Speakers

Synthetic Personas: What to Trust, What to Test, What to Avoid

Arnie Guha  (Partner, Chief Innovation Officer, Phase 5)
Date: Wednesday, October 7
Time: 3:40 pm - 4:10 pm
Pass Type: All Access - Get your pass now!
Session Type: Summit (All Access Required)
Track: Summit | AI in Action
Vault Recording: TBD
Arnie Guha

Synthetic personas are rapidly moving from experimental tool to mainstream offering—appearing in nearly every CX, insights, and marketing pitch deck. But while the term is everywhere, the definitions—and the implications—remain unclear.


What exactly are you being sold? A model of your customers? A simulation of them? Or something in between?


This session cuts through the noise. Rather than taking a position for or against the technology, we offer a practical framework for understanding and using synthetic personas responsibly—grounded in how real organisations make decisions, manage risk, and generate insight.


You’ll learn how to distinguish between different types of synthetic personas, where they genuinely add value, where they require careful validation, and where they can quietly lead organisations astray.


Structured around three simple but powerful questions—what to trust, what to test, and what to avoid—this session equips CX, Marketing, and Insights leaders with the clarity and discipline needed to navigate one of the most rapidly evolving areas in customer understanding today.

Takeaway

  • A clear, working definition of synthetic personas—and how they differ from traditional personas and research
  • A simple taxonomy to identify what kind of solution you are actually being sold
  • Three common failure modes—and how to recognise them before they impact decisions
  • Five critical questions that will transform how you evaluate vendors and internal use cases
  • A practical diagnostic to assess whether your organisation is ready to deploy synthetic personas responsibly
  • Clear guidance on:
    - What to trust synthetic personas for
    - What to test before acting on their outputs
    - What to avoid to prevent costly missteps