What does exceptional design mean when the experience itself is no longer entirely predetermined?
The technology industry has given us enduring design and leadership principles that help us set direction, build strong teams, and create products people value. They give us a shared language for what good looks like. But as AI-native systems increasingly shape experiences in real time, I've been questioning whether the way we apply them is keeping pace with the experiences we are creating.
As software begins generating interfaces, adapting workflows, and acting on people's behalf, we're moving beyond experiences that can be fully designed in advance. Which principles should remain constant, and how must we evolve the way we put them into practice?
I was recently brushing up on Amazon’s leadership principles, particularly Invent and Simplify. One line has always resonated with me: “We accept that we may be misunderstood for long periods of time.” I recognize that part of leadership. Sometimes we see a possibility before others do, connecting what is changing with what is missing. Moving that vision forward takes conviction, especially when stakeholders are skeptical.
Leadership requires holding that space: giving an idea time to develop, making it tangible, and staying open to what other people and new evidence reveal. That feels especially relevant to AI-native products.
Where we once anticipated people's needs and designed defined workflows, experiences can now adapt to individual tasks, contexts, and behaviors. We can no longer specify every path in advance, but we remain responsible for the quality of the experience.
For me, exceptional means solving a meaningful problem with clarity, coherence, and care. The experience earns trust, gives people appropriate control, and feels distinctive. Its emotional character fits the situation: confidence in complex work, relief when a difficult task becomes manageable, or delight in discovering a possibility.
Ultimately, its value is demonstrated through outcomes.
What enduring principles still get right
I revisited the published principles of Amazon, Apple, Google, Microsoft, Airbnb, and IBM to understand what continues to matter as experiences become increasingly adaptive. While these frameworks span leadership, product development, and design practice, they share enduring commitments to understanding people, pursuing quality, building trust, and taking responsibility for outcomes.
| Company and framework | Emphasis that stands out in this comparison | Connection to my own thinking |
|---|---|---|
| Amazon Leadership Principles | Customer commitment, invention, ownership, judgment, and delivery | Hold a future direction through uncertainty and make progress tangible |
| Apple Human Interface Guidelines | Purpose, agency, responsibility, familiarity, flexibility, simplicity, craft, and delight | Give a product a clear reason to exist and make every interaction express that intent |
| Google Ten things we know to be true | User focus, speed, focused excellence, and continued improvement | Respect people’s attention and keep improving what they can accomplish |
| Microsoft Inclusive Design | Identifying exclusion and learning from diverse experiences | Begin with real people in real contexts and build accessibility into the work |
| Airbnb Design Language System principles published in 2016 | Unified, universal, iconic, and conversational experiences | Combine a coherent system with a recognizable product identity |
| IBM Enterprise Design Thinking | User outcomes, repeated reinvention, and diverse empowered teams | Connect discovery, experimentation, and delivery around a shared outcome |
The common foundations are clear: understand people, make deliberate choices, pursue quality, keep learning, and take responsibility.
Inclusion, agency, coherence, and emotional experience also recur.
These commitments align closely with my own.
The foundations I would carry forward
I would carry forward Amazon’s Customer Obsession, Ownership, Earn Trust, Learn and Be Curious, Insist on the Highest Standards, and Deliver Results.
Apple’s attention to purpose, craft, agency, and delight also resonates deeply with me.
Microsoft’s inclusion methodology and IBM’s focus on user outcomes reinforce commitments I would preserve.
Great experiences require a strong point of view about what should exist, how it should work, and why someone would choose it. In an AI-native world, that point of view must guide not only product decisions, but the systems making decisions as the experience unfolds.
Evolving the principles for AI-native experiences
The commitments remain relevant.
What is changing is where and how we put them into practice. Our vision, standards of craft, and human judgment now need to guide the systems making decisions as an experience unfolds. That expands how we put these commitments into practice.
Commitment 1. Preserve a differentiated product vision.
A differentiated product vision begins with a deep understanding of people, the problems they face, and a distinctive approach to solving them. When experiences adapt to individual needs and contexts, that vision must guide more than a predefined interface. The company’s unique brand, core values, and approach to solving the problem need to inform the rules and guardrails shaping the experience. Those choices should influence what the system prioritizes, how it communicates, and how it supports people. The experience may change, but what makes it valuable and worth choosing should remain recognizable.
For example, a company expanding into value-based care might guide its system to emphasize continuity, prevention, and coordination alongside the needs of each visit. That strategic direction becomes recognizable in what the experience prioritizes.
Commitment 2. Make craft part of the system.
When an experience can generate itself, craft must be embedded in the rules designers define to guide how AI composes and renders it. Those rules carry design judgment about information hierarchy, interaction, accessibility, and visual coherence, helping ensure that what the system produces is usable, useful, and thoughtfully crafted.
Consider an interface that adapts to a user's task and level of urgency. The layout may change, but information hierarchy, accessibility, and product identity should remain cohesive. Qualities can no longer depend on a designer reviewing a fixed screen; rather, they need to inform the rules and evaluations guiding what the system generates.
Commitment 3. Build confidence that earns trust.
Trust develops as people gain confidence that a system is getting things right and can recognize when it may be wrong. That confidence must be built into the experience: making the basis for its conclusions understandable through sources and evidence, acknowledging uncertainty, and asking for more input when needed. People need meaningful control, with clear ways to question, correct, or intervene at any point. The system must also be designed to learn from feedback and corrections, improving its reliability over time.
For example, an AI-generated patient summary should connect its conclusions to supporting evidence and original source data, allowing clinicians to inspect the details and reach their own conclusions. The summarization layer and the trust layer must be designed into the system together, making it seamless to move from an overview to the evidence behind it.
Commitment 4. Extend accountability beyond delivery.
As experiences evolve, human judgment must remain part of how systems are guided, evaluated, and improved. Governance defines who is accountable, which decisions require human review, and when the system must pause or escalate. Keeping people in the loop requires clear responsibilities, meaningful opportunities to intervene, and processes for turning real-world evidence into improvements to the system.
For me, craft belongs inside the guardrails themselves.
They should guide the system toward clarity, coherence, usefulness, and emotional appropriateness while preserving what makes the product distinctive.
Human judgment shapes those conditions and continues to improve them. Systems can make many decisions within tested boundaries, with people involved where the context, consequences, or limits of established guidance require their expertise.
Much of my work explores how to make this practical: connecting differentiated product vision with design intelligence, experience governance, and evidence from what the system produces.
These connections point to a broader evolution in design leadership, one that extends beyond the experiences we create to the intelligence and systems that shape them.
We have to define exceptional clearly enough to guide a system, while staying close enough to people to recognize when it falls short.