Exploring AI Simulation for Accelerating Design: A New Approach to UX in the Age of AI

Join Alex Les, VP of Data Science and AI at Huge Design and Technology Company, as he explores how AI simulation can accelerate design by empowering designers to work with invisible users and turn data artifacts into active participants in the design process.

  • 1. Alex Les is the VP of Data Science and AI at Huge, a design and technology company.
  • 2. The talk focuses on "invisible users" and "invisible interfaces," discussing the current state, future potential, and an idea to bridge the gap in UX and AI.
  • 3. There is a trust gap in AI, with only 32% of US adults and 44% of adults globally trusting AI according to recent research by Edelman (Dec 2024).
  • 4. The reason for this lack of trust is "AI slop," where websites, products, or interfaces provide incorrect information due to AI failures.
  • 5. Users can talk to machine learning models in natural language with the help of Genai, which is a UX revolution according to Cassie Kosakov.
  • 6. Don Norman of Neielson Norman group emphasized simplicity and efficiency as well as invisible interfaces that feel seamless and intuitive for users.
  • 7. The idea is to use AI to design interfaces that feel like magic through accelerated need finding, a process proposed by Alex Les.
  • 8. The new approach to the design life cycle would involve empowering designers to work with "invisible users" in the form of AI simulation.
  • 9. Designers can collect data artifacts during need-finding and turn them into active participants in the design process using AI simulation.
  • 10. This new approach aims to deliver better design by providing designers their own mini feedback cycle to work with AI simulation.
  • 11. The accelerated need-finding process includes defining audiences, mapping intentions, identifying tasks, conducting analysis, refining insights, and developing design alternatives.
  • 12. In the era of AI simulation, there are key differences in some components and workflow for need finding, such as starting with data representing target audiences.
  • 13. At Huge, they have a data platform called "live" that contains demographic, psychographic, and contextual datasets to simulate audience behaviors in real-world scenarios.
  • 14. Using the intelligent twin methodology, designers can create simulations of user behaviors and desired outcomes for specific tasks and interfaces.
  • 15. AI simulation allows for high-level understanding across an entire category as well as nuanced insights into specific areas of the experience.
  • 16. The audit process can generate a design brief that allows human teams to focus on solving pain points distracting users in a specific category, improving the overall design process.
  • 17. AI simulation enables faster and better data gathering, helping design teams create websites and interfaces that restore and repair the trust gap between users and AI.
  • 18. In 2025, there has been progress with tools like Anthropic's MCP protocol, making it easier to create new designs by turning prototype components into actual code components in popular frameworks.
  • 19. Limitations of AI simulation include reproducibility, which can be addressed by standardizing parameters and applying a test-and-control methodology.
  • 20. AI simulation should be used alongside human teams, with considerations for different industries, geographies, and domains to isolate the strengths of intelligent twins for design need finding.
  • 21. The future of UX and AI lies in creating better websites, mobile apps, and interfaces that emphasize clarity, simplicity, and trustworthiness.
  • 22. Users don't need more chatbots that don't work; they need better surfaces that provide a seamless user experience.
  • 23. The proposed new approach to design aims to help teams gather insights in a smarter, faster, and better process using AI simulation.
  • 24. By empowering designers with AI simulation tools, we can bridge the trust gap and create websites and interfaces that restore and repair user trust in AI.

Source: AI Engineer via YouTube

❓ What do you think? What are the implications of using AI simulation to accelerate the design process, and how might this impact the relationship between users and technology? Feel free to share your thoughts in the comments!