How we restructured Airtable’s entire org for AI | Howie Liu (co-founder and CEO)
How we restructured Airtable’s entire org for AI | Howie Liu (co-founder and CEO)
How we restructured Airtable’s entire org for AI | Howie Liu (co-founder and CEO)
In a conversation with Howie Liu, CEO of Airtable, we explore how the company is redefining its approach in the age of AI. Following a wave of external skepticism and internal transformation, Airtable has embraced a new operating model that blends rapid innovation with deep technical engagement. Liu discusses the strategic shifts that have allowed the company to not only survive but thrive, including his personal return to hands-on product development and a reimagined team structure designed for speed and agility.
Howie Liu details Airtable’s transformation after a viral tweet questioned the company’s viability. Faced with external criticism and internal challenges, he restructured the company into fast- and slow-thinking teams inspired by Daniel Kahneman’s framework, enabling rapid AI feature development. Liu also embraced the role of an 'IC CEO,' actively coding and using AI tools daily, which has helped drive over $100 million in free cash flow. He emphasizes the importance of experimentation, urging employees to cancel meetings and play with AI tools to discover new possibilities. The company has also shifted its product philosophy, integrating AI more deeply into its platform through features like Omni and encouraging cross-functional skill development. Liu argues that in the AI era, CEOs must stay close to the product, traditional role boundaries must blur, and evals should follow, not precede, creative exploration.
04:10
04:10
A viral tweet falsely claimed Airtable was dead and overvalued.
08:08
08:08
Every software product needs to be refounded in the Gen AI era.
14:01
14:01
New LLM map-reduce capability offers strategic value akin to a smart chief-of-staff analyzing data.
16:27
16:27
The speaker cut one-on-one meetings to focus on timely topics and real insights.
29:16
29:16
ChatGPT's rapid adoption demonstrated the power of experiential AI learning.
32:57
32:57
Cursor can perform complex app-building tasks beyond basic code autocompletion
35:04
35:04
Airtable's no-code components enable agents to build apps more effectively.
40:20
40:20
Using AI tools like Runway and Sesame helps understand GPT-5 and other models better.
49:19
49:19
Prototypes allow for testing in realistic scenarios and getting a feel for the product.
1:01:50
1:01:50
Implementing deep research powered by ChatGPT for field agents
1:04:10
1:04:10
Evals are most useful after converging on the basic product form and use-cases.
1:10:51
1:10:51
Employees need to become more full-stack and outcome-oriented across all functions.
1:20:32
1:20:32
Being involved in product details is invigorating and leads to non-incremental outcomes.
1:22:21
1:22:21
Entrepreneurs should remain close to what they love as their company scales.
1:28:07
1:28:07
Modern tools like Cursor have minimized the gap between the builder and the final product.
1:30:00
1:30:00
The Three-Body Problem series is mind-expanding and gets better after one and a half books
