From UX to HAX Design: Designing for the Age of AI

I came into UX design just before the first internet bubble burst. Websites in those days were genuinely difficult to use. Navigation was an afterthought, information architecture was not yet a phrase most teams knew, and the technology was still finding its footing. Pages loaded slowly, layouts broke across browsers, and the notion that a user’s experience deserved deliberate attention was, for many organisations, a novel idea. The problems were obvious, the impact of good design was immediate, and there was a real sense that the discipline was being built in real time.

What followed was remarkable. UX designers did not just improve products; they changed how an entire generation of people related on technology. We established the conventions that made the web navigable: consistent navigation, clear hierarchy, accessible interfaces, responsive layouts. We pushed for user research when nobody wanted to slow down for it. We fought for accessibility when it was treated as optional. We introduced design systems that brought coherence to digital chaos. We shaped how billions of people shop, communicate, learn, bank, and access information. That was not a small thing. That was a profession proving its worth at a human scale.

Then, in the early 2020s, something broke. Not a feature, not even a product. The interaction model itself changed. And we as UX Designer were unable to handle this behaviour in our daily challanges.

The Breaking Point

The arrival of large language models and generative AI did not simply add a new capability to existing products. It invalidated the foundational assumption that UX was built on: that systems behave predictably. As Jakob Nielsen observed, UX until now had been grounded in Human-Computer Interaction: a set of tasks performed by a user to reach an expected output. Artificial intelligence replaced that with something entirely different: intent-based outcome specification, where the user expresses the final output they intend rather than completing a defined sequence of steps. Medium

Suddenly, the user journey map could not account for a system that interprets rather than executes. The usability test could not pin down an output that changes each time. The design specification could not define states for a model that adapts continuously. AI practitioners began describing it as “the most ambiguous space I’ve ever worked in, in my years of working in design… There aren’t any real rules and we don’t have a lot of tools.” Microsoft That was not a junior designer venting. That was the field admitting, collectively, that it was not prepared. I know I was not prepared for what came, at all…

The rules you built your career on no longer apply. And the products being designed without you are already in the hands of billions of people.

And the stakes are not abstract. Unclear user expectations about the set of supported tasks or domains can lead to disappointment, product abandonment, and even harms. Microsoft Research on human-AI experience remains fragmented, highlighting the need for a unified overview of current practices, challenges, and opportunities. arXiv Meanwhile, the rise of generative and autonomous agents marks a fundamental shift in computing, demanding a rethinking of how humans collaborate with probabilistic, partially autonomous systems. arXiv Someone is designing those systems right now. The question is whether designers are in the room.

The Resourceful Response

The first steps towards AI Centered UX

Human-AI eXperience Design (HAX) is not a rebranding exercise. It is the field’s honest attempt to build a new practice from real observed problems. HAX aims to create AI systems that are user-friendly, trustworthy, ethical, and beneficial, covering the broad spectrum from narrow AI to generative AI. IxDF Where UX asks “how does the user complete this task?”, HAX asks “how does the user collaborate with a system that is adaptive, sometimes wrong, and continuously learning?”

The shift is practical, not philosophical. In classical UX, error states are edge cases. In HAX, they are designed-for certainties. The AI will hallucinate. The AI will misunderstand context. It will, over time, change its behaviour as it adapts. These are not failures of design; they are properties of the medium. A HAX designer must account for the entire relationship between a user and an AI system, not merely the task at hand.

This required new tools. And practitioners built them.

The Work Being Done

Microsoft’s HAX Toolkit Project started as a collaboration between Microsoft Research and Aether, Microsoft’s advisory body on AI Ethics and Effects in Engineering and Research, offering practical tools for creating human-AI experiences, each grounded in observed needs, validated through rigorous research, and tested with practitioner teams. Microsoft Its foundation is a set of 18 generally applicable Guidelines for Human-AI Interaction, introduced at the 2019 ACM CHI Conference and validated through a user study with 49 design practitioners testing the guidelines against 20 popular AI-infused products. Microsoft The guidelines are divided into four groups: initially, during interaction, when the AI system gets something wrong and needs to be redirected, and over time. Microsoft Alongside them, the HAX Design Patterns describe flexible, actionable solutions to recurring human-AI interaction problems, derived by analysing hundreds of examples in everyday AI products. Microsoft

The field is not waiting for a consensus. It is being built in parallel, by researchers, practitioners and institutions who recognised the same gap and started filling it from different directions.

Microsoft is not alone. Google’s People + AI Guidebook is a set of methods, best practices, and examples for designing with AI, based on insights from over a hundred experts, with Google covering mental models, explainability, trust, and graceful failure. It has been adopted by practitioners across the world. IBM Research produced a set of design principles for generative AI applications, developed through an iterative process involving literature review, practitioner feedback, and validation against real-world products, ACM Digital Library addressing the specific characteristics of generative systems that earlier frameworks were not built to handle. Carnegie Mellon University’s Human-Computer Interaction Institute created an AI Brainstorming Kit providing a structured approach to designing AI-driven solutions that are both technologically feasible and user-centred, UX Design sitting earlier in the process than most toolkits and addressing problem framing before any design decisions are made.

At the research level, JetBrains’ HAX research team spans three core areas of design, impact, and quality, going beyond traditional UX research and A/B testing by leveraging behavioural sciences and experimental mixed-method studies. JetBrains And HAXD 2025, the International Symposium on Human-AI Interaction and Experience Design, addresses the full spectrum of user outcomes: usefulness, usability, trust, cognitive workload, satisfaction, safety, fairness, accessibility, and long-term societal impact, Intelligent-systems signalling that HAX is no longer a side conversation within HCI. It is becoming a discipline in its own right.

None of this happened in a workshop. It happened because practitioners and researchers looked at a broken situation and built their way out of it. Sound familiar?

This Is Your Moment Again

The Adventure Continues as UX Professionals

We have been here before. Not in the details, but in the shape of the challenge. A technology arrives faster than the design community is ready for. The rules do not exist yet. The stakeholders do not fully understand what they are building. Users are left to navigate something that was not designed with them in mind. And then UX designers step in, make sense of it, and quietly reshape how the world interacts with something new.

We did it with the early web. We did it with mobile. We did it with touch interfaces, with voice, with responsive design. Each time, the tools did not yet exist. Each time, we built frameworks, established conventions, and dragged an entire industry toward something more human. Each time, we redefined what the profession meant and what it was capable of.

That cycle is happening again. And for experienced UX designers, this is not a threat to your career. It is the most significant opportunity your career has ever presented.

The transition to HAX requires the same mindset expansion that every previous shift demanded. Your skills in empathy, research, prototyping, and testing are not obsolete. They are the foundation every one of these frameworks is built on. What changes is the object of design. You are no longer designing a defined journey through a deterministic system. You are designing a relationship with a probabilistic one. Trust becomes a design material, something that must be earned, communicated, and repaired when broken. Ambiguity becomes the default state, not an inconvenience to be resolved before work begins.

The toolkits from Microsoft, Google, IBM, and Carnegie Mellon are not waiting for a new generation of designers trained in machine learning. They are waiting for experienced designers who understand humans. That means you.

Stay hungry…

Please go through the Microsoft HAX Guidelines and work through the Workbook with your team. Explore Google’s People + AI Guidebook. Study IBM’s generative AI principles. Understand what Carnegie Mellon’s Brainstorming Kit is doing at the problem-framing stage. These are not theoretical documents; they are working tools, built by practitioners, for practitioners. By UX Professionals for UX Professionals…

We have fixed the broken websites before. We brought structure to the chaos of early digital products. We made the internet usable. We shaped how a generation of people live their digital lives. This is that moment again, only bigger, faster, and with considerably higher stakes.

Every dark pattern, every confusing interface, every product that harmed instead of helped, happened because design was absent from the room. AI is now in every room. But AI can’t exist without UX Professionals, when it’s made for humans, by humans…

The users who will be confused, misled, or even harmed by poorly designed AI experiences need someone in the room who knows better. You know better.

Where are you?

In retrospect. Steve Jobs closed his 2005 Stanford commencement address with two words borrowed from the back cover of the Whole Earth Catalog: “Stay Hungry. Stay Foolish”. He was not talking about desperation or naivety. He was talking about the kind of insatiable curiosity that refuses to accept that the current answer is the final one. That quality built the personal computer industry. It built the web. It built the design profession you chose. It is the same quality that HAX Design now demands. The designers who will define this era are not the ones who already know how AI works. They are the ones who cannot stop asking what it means for the people who have to live with it. Stay curious. Stay restless. The most important design problem of your generation is waiting.

HAXDesign #UXDesign #HumanAIExperience #DesignForAI #MicrosoftHAX #ResponsibleAI #GenerativeAI #HumanCenteredAI #FutureOfDesign #UXCommunity #AIEthics #Design