Airtable Launches Superagent, Its First Standalone AI Agent Product

Airtable is launching an entirely new product line at a moment when its valuation looks very different from the peak of the last tech cycle — and CEO Howie Liu says that’s precisely why the move makes sense.

Airtable was valued at $11.7 billion during the zero-interest-rate boom of 2021. Today, shares trading on secondary markets value the company at closer to $4 billion. Liu argues that while the drop hurt investor paper returns and employee stock options, it didn’t damage the underlying business. Airtable has raised $1.4 billion in total funding, retains roughly half of that capital, and, according to Liu, continues to generate cash.

Rather than retrench, the company is expanding. Airtable is introducing Superagent, an AI agent product that operates independently from its core no-code platform and represents the company’s first standalone product in its 13-year history. Liu says Superagent could eventually grow large enough to rival Airtable itself.

The launch reflects both Airtable’s long-term direction and the current state of the software industry, where nearly every major player is racing to prove it can deliver capable AI agents. Airtable, which employs more than 700 people and serves more than 500,000 organizations — including about 80% of the Fortune 100 — is not a struggling startup. It’s a mature company making a deliberate architectural bet.

Superagent is built around what Liu calls “multi-agent coordination.” Instead of a single assistant working through tasks one at a time, the system orchestrates multiple specialized agents in parallel. “You’re not prompting an AI,” Liu said. “You’re orchestrating a team.”

When a user asks Superagent a complex question — such as whether to expand an athleisure brand into Europe — the system first creates a research plan, identifying key dimensions to investigate. It then deploys agents simultaneously to examine areas like financials, competition, management, and market conditions. The final output combines those threads into a single, structured deliverable.

That deliverable isn’t just text. Superagent produces interactive analyses with visualizations, demographic data, competitive maps, and timelines that users can explore and filter. Liu described the shift as fundamental. “What if every person could have New York Times–quality data visualization built for every task they have?” he said, arguing that high-quality interactive outputs becoming the default is a significant change from earlier generations of AI tools.

Liu draws a technical distinction between Superagent and many other products labelled as AI agents. He points to Anthropic’s Claude and Manus — a research-focused agent platform being acquired by Meta — as the only systems he considers competent, long-running agents. Many other offerings, he argues, are closer to predefined workflows with embedded AI calls than to autonomous systems that can backtrack and adapt.

That claim will need to hold up in a crowded market. OpenAI kicked off 2025 with new agent-building tools, and companies like Notion, Harvey, and many others have added agent features. In a space full of competing claims, Superagent’s differentiation will depend on real-world performance.

In a blog post announcing the product, Liu outlined several use cases. Asked to evaluate Google as a three-year investment, Superagent produces a structured analysis that cites earnings calls, assesses competitive threats from OpenAI and Anthropic, and highlights overlooked risks. Asked to brief a user before pitching Wells Fargo, the system pulls together regulatory context, recent AI investments, and specific operational pain points. To do this, Superagent draws on premium data sources including FactSet, Crunchbase, SEC filings, and earnings transcripts.

The product caps a broader repositioning of Airtable as what Liu calls an “AI-native platform.” Last fall, the company hired David Azose — formerly the engineering lead for ChatGPT’s business products at OpenAI — as chief technology officer. Around the same time, Airtable acquired DeepSky (formerly Gradient), an AI agent startup that had raised $40 million in funding. Superagent will be run semi-independently by DeepSky’s founding team.

Pricing was still being finalized as of last week. Still, Liu suggested a familiar AI SaaS model: around $20 per user per month at the entry level, scaling up to roughly $200 for heavy users, with generous inference allowances. “We’re not trying to optimize for profit margin right now,” he said.

Whether Superagent becomes the massive opportunity Liu envisions remains an open question. Competition is intense, and customers may care more about cost and speed than about philosophical distinctions between “real agents” and advanced workflows.

Still, Liu views the launch as a calculated bet rather than a defensive move. After seeing $7.7 billion in valuation evaporate on paper while retaining most of the company’s actual capital, he says Airtable has flexibility. He has framed the lower valuation internally as a recruiting advantage, telling employees they’re receiving equity priced far below the 2021 peak, with significant upside if the company’s bets pay off.

Asked whether Superagent could ultimately outgrow Airtable, Liu doesn’t dismiss the idea. Airtable, he said, will likely remain larger in the near term. “But I also like being able to bet on Superagent,” he said. “Optionality is a good thing.”

Liu described the strategy as a form of “wartime” leadership — a term he once avoided but now embraces. For him, it means adapting quickly rather than defending the status quo. “Being very fast on the draw to be able to adapt,” he said, is “the most value-creative way to run things right now.” He added, “It’s also the most exciting way to do things.”