Intercom on Product: How to design standout products in an AI world

Generative AI is revolutionizing technology so fast that it already seems unthinkable to design a product without considering where AI will fit in.

LLMs are transforming the way we use our products, both hardware and software, and designers need to think about how they can evolve their methods and grasp the unique opportunities that come with these advanced systems.

In this week’s episode of Intercom on Product, our VP of Design, Emmet Connolly, talks with Staff Product Designer Molly Mahar about her recent article on the topic. They dive into the nuances of designing for AI, exploring their experiences and learnings so far – as well as offering valuable advice for designers looking to navigate this emerging field.

“Designing for a probabilistic system requires us to embrace surprises and design around them”

Here are the top insights from their discussion.

• Expect the unexpected when it comes to AI: Working with AI means moving away from deterministic systems, where outcomes are clear-cut and reliable, to the more dynamic terrain of probabilistic systems. For designers, this introduces a level of unpredictability and complexity and calls for a design strategy that’s not just robust, but also flexible and adaptable. Molly sums it up when she says, “Designing for a probabilistic system requires us to embrace surprises and design around them.”
• Data quality is the foundation of reliable AI systems: As the saying goes, “Garbage in, garbage out.” The effectiveness and reliability of AI systems rely on the quality of the data they’re trained with. Designers should prioritize data quality to optimize product performance and ensure user satisfaction.
• Use realistic prototyping to elicit valuable user feedback: In the AI landscape, taking an iterative approach to prototyping becomes even more important. Wherever possible, make the prototype experience as realistic as possible by choosing a simple visual design but employing actual data. This allows designers to capture essential user feedback and evaluate the effectiveness and user-friendliness of the output.
• Test, test, and test again: Designers need to anticipate all possible system failures and set up backup plans to deal with them. If the live version throws a curveball that catches designers off guard, it’s a clear sign that the testing phase needs to be more robust.
• Prioritize winning your users’ trust: By focusing on building trust through transparency and simplicity, designers can develop AI experiences that make users feel comfortable and confident. Establishing trust means creating consistently positive experiences and offering visibility into the way the system works and the data it works with.

For designers navigating these new challenges, it’s important to stay curious and adaptable, and to keep the user at the center of every decision. By understanding the importance of data quality, leveraging iterative prototyping, planning for system failures, and building trust with users, designers can embrace the new age of UX and become a success story of the AI revolution.

If you enjoy our discussion, check out more episodes of our podcast. You can follow on Apple Podcasts, Spotify, YouTube or grab the RSS feed in your player of choice.