Howie Liu on Airtable’s Enterprise Strategy and AI Adoption

In this August 2023 interview, Airtable founder and CEO Howie Liu tells Harry Stebbings that he wishes he had connected product strategy and go-to-market earlier. Product-market fit matters, he says, but different products require different distribution and sales models.

Airtable was built for real-time collaboration, but its initial templates included both individual and team use cases. In hindsight Liu would have guided adoption toward larger teams sooner, because those uses could support more sales and marketing approaches. This is his retrospective judgment, not proof that the alternative would have performed better.

For a horizontal product with unclear use cases, he recommends deeply solving a few specific workflows while keeping the platform flexible enough to expand. He contrasts that with products such as chat or file sharing, whose general use is more immediately understood.

Liu thinks generative AI could ultimately have effects comparable to or greater than cloud computing, because it may assist many kinds of knowledge work. This is a prediction made in 2023 by the leader of a company building AI features, not a demonstrated economic outcome.

He says enterprise customers were still learning what AI could do and where it could fail. Hallucination, accuracy and the ability to cite sources matter especially for high-stakes uses such as benefits or financial information. Interest in a model does not establish that it is safe or suitable for a particular workflow.

He argues that consumer applications exposed many people to AI before formal enterprise deployments, potentially making the technology more accessible. His examples include people outside technology experimenting with research. He also acknowledges that outputs can be wrong and require scrutiny.

Stebbings raises employees’ fears that AI tools may replace them. Liu calls the economic consequences serious and says changes could arrive faster than societies adapt. He sees a possible future in which productivity reduces costs and expands demand and employment, but explicitly says that outcome is not guaranteed.

In his view, enterprise adoption in 2023 was still at an education stage. He lists privacy, trust in hosted providers, training-data provenance, copyright, accuracy and safety as practical barriers. His comments about particular provider capabilities are historical, not a current product or legal assessment.

Liu expects services firms and systems integrators to help enterprises with technical choices such as retrieval, embeddings, document handling and citations. He does not know how much value service providers versus application companies will ultimately capture. Some firms can build internally, but he says doing so requires substantial expertise.

He does not see AI as a simple winner-takes-all contest between incumbents and startups. Existing vendors may add value to products customers already use, while newer companies might create experiences for different users or workflows. This is his market thesis rather than a forecast of specific winners.

Liu says Airtable can move quickly, but acknowledges a tension: building a new standalone product may be faster than integrating features into a mature one that must satisfy existing customers. Security, stability and service expectations can add work, while integration can compound the value of the current product.

Asked whether enterprises would adopt AI en masse in the following year, he says no particular year is guaranteed. He sees a continuing learning loop: try lower-stakes use cases, observe results and iterate. Airtable’s AI field is his example of a building block for combining model outputs with existing workflows.

Potential uses he discusses include synthesizing customer feedback, generating product ideas and summarizing progress. He says Airtable wants to test specific use cases with customers rather than assume it knows all the important ones. His description is a company product vision, not an independent performance evaluation.

Moving from product-led growth to enterprise sales requires aligning product value with the sales motion. Liu argues that a tool used by many individuals can be difficult to sell to a senior buyer if it cannot show a meaningful team or organization-level benefit.

Enterprises may seek to reduce overlapping tools and the burden of supporting many vendors. Liu says a supplier needs a clear business case for an executive buyer, rather than relying only on the number of individual users. This is his sales judgment, informed by Airtable’s commercial interests.

He thinks broadly applicable but relatively interchangeable productivity tools are especially vulnerable to bundles. A CIO may choose one suite even if some employees prefer individual alternatives. He contrasts that with a product that can demonstrate differentiated business impact.

Liu positions Airtable as an application platform for processes such as marketing and product operations. He claims it can be faster to build on than some alternatives and create measurable business value. Those competitive and ROI claims come from Airtable’s CEO and are not independently verified in the conversation.

He says leading an enterprise sales conversation with AI features alone has not reliably closed deals. Curiosity can open a meeting, but buyers need a specific problem, business justification and budget. He favors discussing how AI might improve an actual workflow rather than treating AI itself as the value proposition.

Stebbings notes that a large corporate logo can represent only a small experiment. Liu considers a roughly $1 million annual contract a useful threshold for a truly significant enterprise relationship, while acknowledging that $250,000 or $500,000 can also be meaningful. This is his rule of thumb, not a universal standard.

Liu recalls Airtable reaching its first approximately $1 million annual contract around 2019, several years after launch and after substantial organic adoption. He says large organizations already had thousands of active users. The timing and usage figures are his recollection.

The guests discuss buyers demanding evidence of value as budgets tighten. Liu says companies are reviewing unused seats and overlapping software after a period of rapid digitization. He regards that review as understandable rather than a surprise.

An active-to-paid-seat ratio is only a shallow measure, he says: heavy usage does not necessarily improve business outcomes. He recounts a CIO observing that vendors’ claimed time savings could add up to more than a working week. Liu wants post-sales teams to demonstrate actual results rather than simply count activity.

As an angel investor, Liu says it is too early to put everything into one AI investment because he does not know which companies will win. Seeing other founders closely has also taught him that businesses look easier from the outside than they are.

Discussing a later venture round, he says investors were attracted by a large potential market and accelerating adoption. He recalls revenue growing from roughly $1 million to $10 million in a little over a year. These are historical company figures and a founder’s account, not guaranteed investment returns.

He rejects the idea that finding product-market fit makes everything afterward easy. Building and scaling the business introduces new challenges in product, operations and distribution. Successful early adoption is the beginning of that work, not its completion.

Airtable spent about two and a half years building before launch. Liu wishes he had combined long-term patience with clearer milestones and accountability along the way. He mentions a reported LinkedIn planning example but says he does not know whether the story is entirely accurate.

Asked about pressure from the company’s previous valuation, he says he focuses on durable growth and execution because valuation is an outcome affected by market multiples. Stebbings cites approximately $12 billion while the video title gives $11 billion; these are historical figures in the program, not current valuations.

Liu believes efficient, sustained growth could eventually support the prior valuation and more. That is his aspiration as founder and shareholder, not a guarantee. He acknowledges the work of achieving it is difficult and that market multiples are outside the company’s direct control.

He says he cares less about personal legacy than about doing right by employees and shareholders. Over time he has learned that building a good product and building a durable company are different challenges. The interview closes with mutual thanks.