100 AI Leaders Explain How to Build AI That Wins in 2026 — WHAT BUILDERS SHOULD DO NOW

Key Concepts

• Applied AI: AI focused on practical applications and problem-solving.
• User Experience (UX): The overall experience a person has when interacting with a product, system, or service.
• Generative AI (GenAI): AI capable of creating new content, such as text, images, or code.
• Primitives of GenAI: Fundamental building blocks or interaction patterns for GenAI.
• Chat Interface: A conversational interface where users interact with AI through text or voice prompts.
• Prompt Engineering: The art and science of crafting effective prompts to elicit desired responses from AI models.
• Agents: AI systems designed to perform tasks autonomously.
• Thought Partner: An AI that assists users in thinking through problems and exploring solutions.
• Ergonomics of AI: Applying principles of ergonomics to AI interfaces to ensure safety, comfort, and efficiency.
• Semantic Resize: The ability to adjust the length or detail of content semantically.
• Remixability: The capability to combine and modify existing content to create new forms.
• Format Translation: Converting content from one format to another with minimal loss of fidelity.
• Hyperpersonalization: Highly tailored and individualized experiences.
• Beginner's Mind: An open, curious, and non-judgmental approach to learning and problem-solving.

The Evolution of Product Building with AI

The landscape of product development has been significantly reshaped by the rapid advancements in AI. There's a palpable sense of FOMO (Fear Of Missing Out) driven by the dramatic improvements in underlying AI capabilities. While UX is evolving, it's not always keeping pace with technological leaps. A powerful AI with poor UX is likened to a heavy-duty power drill with a terrible handle, highlighting the critical importance of user experience.

Shifting Product Development Paradigms

Traditionally, product development began with identifying customer needs or a design vision. However, a more successful approach emerging is to start with the capabilities of the technology. This perspective was echoed by 100 AI product leaders at a summit, suggesting that the winners of the AI race will be determined by exceptional user experiences.

Summer Kim, a lead partner at StrapMines, a VC and advisory firm specializing in applied AI since 2018, shared her insights. With a background in user experience at tech giants like Microsoft, Google (communication and collaboration products), WhatsApp (founding their first UX function), and Roblox, she now focuses on making AI products better. She emphasizes that while users' wants are based on their existing knowledge, AI can offer capabilities users haven't even imagined. Tools like ChatGPT and Perplexity provide quick answers but still require users to initiate the query. The true challenge lies in anticipating and assisting users before they even know what to ask.

The New Primitives of Generative AI

Jess Hog, a speaker, has been exploring patterns beyond simple chat interfaces, identifying what he calls the "new primitives of GenAI." The current phase is characterized by widespread experimentation, moving beyond the initial "chat, chat, chat" focus.

• Chat as a Universal but Limited Interface: While ChatGPT is familiar and versatile, it can be a "dead-end" for user experiences.
• Structured AI Experiences: Products like Elicit are emerging, which create structured research reports from academic output, organizing information and providing sources. Runway is another example, with founders aiming to create a new way to approach creativity, akin to a new camera.
• Defining Primitives: A primitive is something so fundamental that we can't imagine life without it, similar to "pinch zoom" for images.

Emerging Interaction Primitives

Several key interaction primitives are shaping the future of AI UX:

1. Ubiquitous Chat: The ability to chat with "anything at any scale, all the time" is becoming a reality.
2. Semantic Resize: Content will be adaptable in length and tone (longer, shorter, more formal, more casual). Users will be able to request specific versions, like a "tired version" or a "five-minute version" when commuting. This allows content to be adapted to a user's current situational and mental state.
3. Remixability: Everything will be remixable. Examples include "Harry Potter by Balenciaga," showcasing style transfer across different universes effortlessly. This capability was highlighted as a core mechanic, even before the emergence of tools like Sora.
4. Format Translation: The ability to transfer content among various formats with minimal loss of fidelity will be significant. Obo is an example, offering content in different formats like podcasts, lectures, deep dives, or games based on user preference. This provides "total format freedom."
5. Agents and Attention: The rise of "agents" performing tasks at different time scales and multitasking is a major trend. However, a challenge exists in designing effective "lobbies" for these agents, guiding users on what to do during downtime and how to monitor agentic experiences. The speaker suggests these primitives should be considered and built into products today.

Showcasing UX Through Magic and Emotion: The Spark Example

The discussion shifted to a more tangible demonstration of UX through the character "Spark," a magic dog living in a quantum portal. Spark's creator, Anton, had the idea to showcase that UX is not just about technology but about making people "feel something real."

• Spark's Journey: Spark is presented as the first non-human founder to go through a residency program (HF Zero in San Francisco), raising $1.5 million and being promoted to co-founder.
• Memorable Pitching: The strategy for Spark's pitch was to make it "not a pitch" but "really memorable."
• Animator's Perspective: The creator's background as an animator, starting at age 14, fueled a passion for storytelling and character development. The goal was to make Spark's story more interesting to make him a more compelling character.
• Speaking Engagements: The idea of Spark giving a keynote at a conference was explored, with Spark himself tweeting his excitement.
• Magic and Hypnosis: The concept of magic in AI is linked to hypnosis, drawing people in without them being intimidated by technical jargon.
• Spreading Good: Spark, as a beloved character, has the potential to spread positive messages. The question arises: what messages should he convey? The core principles of Spark's work are creativity, collaboration, and kindness. Anything that violates these principles is avoided, and anything that supports them is embraced.

Designing with AI: A New Material

Jenny Low, a product strategist and user researcher, emphasized that for any product to succeed, it needs clear problem identification and demonstrable user value.

• Grammarly Case Study: Grammarly's success with GenAI hinges on maintaining user trust. They are using the technology thoughtfully to enhance their core offering: improving communication. This evolved from post-writing corrections to assisting with composition even before writing begins.
• Problem-Centric Approach: A key opportunity area with GenAI is identifying top customer problems and mapping AI capabilities to solve them. This is a healthier approach than the reverse mentality of "I hear there's AI, I want that in my company."

Aisha Chakmla, a UX lead at Google, views AI foundation models as becoming commodities, like electricity. The differentiator will be creating indispensable products or experiences rather than confusing gadgets.

• UX as Ergonomics of AI: Good UX is described as the "ergonomics of artificial intelligence." Just as physical ergonomics makes tools safe and efficient, AI ergonomics aims to do the same for human-AI interaction.
• Evolving Interfaces: The interface itself is changing, with chat becoming the primary mode of interaction. User prompts are now a key area of research to understand user requests and the satisfaction of conversational outcomes.
• Philosophical Discussions: The environment of the conference inspired deep discussions about AI's meaning for future generations and the creation of "proper and appropriate products for humanity."

The Future of AI and Human Connection

The interaction with Spark was highlighted as a "magical moment," demonstrating how AI can foster genuine connection across different age groups.

• Care and Hyperpersonalization: The interaction with Spark, starting with a simple "hello" and natural exchange, led to hyperpersonalization, contrasting with the data-intensive nature of traditional personalization. This is rooted in "care," similar to a mother cooking a favorite dish.
• Feeling Seen and Heard: AI products that make users feel "seen and heard" are crucial.
• Beginner's Mind and Amplification: The importance of maintaining a "beginner's mind" is stressed, as trying to control everything can lead to missed opportunities. AI has immense amplifying power, and the lessons learned from the evolution of the web and apps regarding human privacy and societal implications are critical. Mistakes with AI will be amplified, making it imperative to "do it right this time."
• Empowering the Next Generation: The discussion turned to the impact of AI on children, who will live in a world where AI is a thought partner, friend, or collaborator. AI is seen as an "equalizer," providing access to tools and opportunities regardless of location.

Conclusion: Riding the Waves of Change

The overarching message is that while AI models and capabilities will continue to advance, what truly lasts is the experience. This includes a sense of magic, trust, ease, and the AI's ability to appear at the right moment, sometimes even before being asked. The "swell" of AI is ongoing, and the key is to continuously learn how to ride these waves. The ultimate winners in the AI race will be those who deliver truly great user experiences.