As product teams race to integrate AI in their products, they often rely on conversational UIs and AI assistants. However, the experiences with them are often painfully slow, the responses are generic and users have to meticulously explain to AI just what exactly they need — over and over and over again.

What’s the state of AI in late 2025? What have we learned about user’s expectations and behaviour patterns? What have we learned about new AI capabilities, reasoning engines and agentic experiences? And what design patterns can we use to help users build trust and confidence to get to better results, faster?

In this session, we'll explore just that — based on actual findings from actual research with actual real-life examples. Roll up your sleeves and buckle up: you’ll leave the session with a toolbox of useful insights that you might need to consider the next time you’ll be working on an AI feature.