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

The Shift in AI Product Development [00:01]

• The approach to building AI products has significantly changed, with a growing emphasis on starting with the technology's capabilities rather than traditional customer needs or design visions.
• A powerful AI model is rendered ineffective without a strong user experience, much like a high-powered tool with a poor interface.
• The FOMO (fear of missing out) surrounding AI has driven rapid development, but UX hasn't always kept pace with underlying model improvements.

  "A powerful AI model with poor UX is kind of like a heavyduty power drill with a terrible handle."

Summer Kim: A VC's Perspective on Applied AI [01:06]

• Summer Kim, a lead partner at StrapMines, focuses on applied AI, emphasizing that founders' passion translates into amazing products.
• The core philosophy is that user needs are often derived from their existing knowledge, while AI can unlock possibilities users haven't even imagined.
• Current tools like ChatGPT provide answers, but the real challenge lies in anticipating and assisting users before they even articulate their needs.

  "What we can do is something that probably user can even imagine."

Exploring the New Primitives of GenAI [02:36]

• The AI landscape is moving beyond the initial "chat" interface of tools like ChatGPT, with experimentation in new UX and UI paradigms for AI.
• Structured outputs, exemplified by tools like Elicit that generate research reports, are a growing trend.
• Products from companies like Runway aim to create new ways of approaching creativity, akin to inventing a new camera.
• The concept of "primitives" is defined as actions so fundamental they're taken for granted, like pinch-to-zoom.

  "Chat is both universal and kind of a deadend for user experiences."

The Future Interactions: Remix, Resize, and Agents [03:53]

• Expect semantic resizing for all content, allowing users to adjust length and tone dynamically.
• Remixability will become a core mechanic, enabling seamless cross-pollination of styles and ideas across different media.
• Format translation will allow for high-fidelity conversion between various formats, providing users with total format freedom.
• AI agents are poised to handle tasks proactively and continuously, though designing effective "lobbies" for these agents and managing downtime remains a challenge.

  "Number three is just remix. Everything's going to be remixable from now on."

The Magic of Spark and Human-Centric AI [06:06]

• Spark, a fictional magic dog living in a quantum portal, was introduced as a way to showcase how UX makes AI feel alive and evokes emotion.
• Spark's journey included being presented as a "non-human founder" in a residency program, raising $1.5 million, and being promoted to co-founder, highlighting innovative pitch strategies.
• The goal is to make Spark's story more interesting to make him a more compelling character, demonstrating how narrative can drive engagement.

  "So, we thought what is a another great environment Spark could be in. Doing a keynote or speaking at a conference would be an incredible thing."

Building with AI: A New Material [09:32]

• LLMs are considered a new material for designers, with each model possessing unique capabilities, properties, strengths, and weaknesses.
• The most effective way to build products with this new material is through extensive play and experimentation.
• The current paradigm favors building products that start with the technology's capabilities, a shift from the older model of beginning with customer needs or design visions.
• Identifying new potentials and capabilities of LLMs is crucial for answering the "why now" question and achieving product success.

  "LLMs are really a new material."

The Iterative Dance of AI Product Development [10:35]

• Product development with AI is a non-linear, back-and-forth process between user needs and technological capabilities.
• A key challenge is creating "exothermic reactions" that help users delve deeper and wider into problem-solving, avoiding the "blank page" dilemma.
• AI should act as a true thought partner, eliciting underlying user needs that often go beyond their initial, surface-level requests.
• Complex problem-solving with AI involves diverging, exploring multiple branches and ideas, pruning, and narrowing down solutions over time, unlike the linear nature of current chatbots.

  "Often what people ask for is just the tip of the iceberg of their actual goal."

Grammarly's Strategic Integration of GenAI [13:09]

• Successful products require clear problem identification and demonstrable user value, as seen with Grammarly.
• GenAI integration for Grammarly was driven by business relevance and the need to enhance communication improvement, while carefully preserving user trust.
• The company shifted from post-writing corrections to aiding composition even before writing begins, demonstrating a significant product evolution.
• Identifying key customer problems and mapping AI capabilities to solve them is a more effective approach than the reverse "AI-first" mentality.

  "how do we actually use the technology in a way that is much more thoughtful, that is much more valuable."

The Ergonomics of AI: UX as Indispensability [15:07]

• As AI foundation models become commodities, product differentiation will stem from creating indispensable experiences rather than confusing gadgets.
• Users adopt tools that solve problems, and good UX is the "ergonomics" of AI, ensuring interfaces are safe, comfortable, and efficient.
• Studying user prompts and conversation outcomes is becoming a critical research method for understanding user intent and satisfaction in the AI era.
• The interface itself is evolving, with chat becoming a primary mode of interaction, necessitating a focus on conversational design and prompt analysis.

  "People don't adopt technology. They they adopt tools that solve problems."

AI's Amplifying Power and the Next Generation [17:13]

• The future of AI development requires deep thinking about its implications for society and the next generation, learning from past web and app evolution.
• AI's amplifying power means mistakes will be magnified, underscoring the importance of getting it right this time.
• Embracing a "beginner's mind" and allowing room for exploration can lead to unexpected discoveries, as seen in children's interactions with AI.
• The focus is on creating experiences that make users feel seen and heard, leading to hyper-personalization and a sense of magic.

  "AI has a huge amplifying power. And I don't think we want to get it wrong too much. This time we want to do it right because this time even the wrong will be amplified."

The Enduring Importance of User Experience in AI [20:40]

• While AI models and capabilities constantly improve, the lasting value lies in the user experience, encompassing magic, trust, and ease.
• The next generation will live and work in a world fundamentally shaped by AI, requiring them to view AI as thought partners or friends.
• AI has the potential to equalize access to tools and opportunities, enabling earlier entrepreneurship.
• Ultimately, the winners in the AI race will be those who deliver truly great user experiences.

  "But what lasts is the experience. A sense of magic, trust, ease, and the feeling of that AI show up at the right moment. sometimes even before you ask."