Launching an entirely new product line while your core business has lost two-thirds of its paper valuation might seem irrational to some. However, Howie Liu, the founder and CEO of Airtable, believes it is the most logical course of action. The company, once valued by investors at $11.7 billion during the peak of the zero-interest-rate era in 2021, now has a secondary market valuation of approximately $4 billion. Despite this, Airtable has raised a total of $1.4 billion, and Liu states the company retains about half of that capital while generating positive cash flow. The dramatic drop in valuation impacted investor returns and employee stock options but did not damage the underlying business. Liu's strategic answer is to introduce Superagent, an AI agent he believes could one day surpass Airtable in significance.

This launch marks Airtable's first standalone product in its 13-year history, signaling both the company's new direction and the current industry-wide imperative: every major software company is now in a race to demonstrate viable AI agents. To appreciate why this move is particularly notable, consider Airtable's foundation. It is a no-code platform that simplifies app development, essentially functioning as a powerful database enabling users to build custom software for their specific workflows. The company employs over 700 people and serves more than 500,000 organizations, including 80% of the Fortune 100. This is not a company in distress but an established business staking its future on a new architectural bet.

Superagent embodies Liu's vision of "multi-agent coordination." This system responds to a query not with a single AI assistant performing tasks in sequence, but with a coordinating agent that deploys multiple specialist agents to work in parallel. "You’re not prompting an AI," Liu explains. "You’re orchestrating a team."

The process works as follows: if you ask Superagent about expanding an athleisure brand into Europe—an example Liu provides—the system first constructs a research plan, identifying necessary investigations and uncovering dimensions you may have overlooked. It then deploys specialized agents simultaneously: one examines financials, another analyzes competitive positioning, and a third reviews management and news coverage. Finally, it synthesizes all findings into a polished deliverable. The output is not a simple text document but an interactive market analysis complete with demographic breakdowns, visually mapped competitive presence, and filterable expansion timelines.

"What if every person could have New York Times-quality data visualization built for every task they have," Liu remarked during a recent Zoom call. "This would have been unfathomable ten years ago, or five years ago, where you don’t get that quality of output - you just get text. But to be able to now get truly extremely high quality, rich interactive outputs as a default format, I think that’s a game changer."

Liu makes a technical distinction between Superagent and its competitors. He cites Anthropic's Claude and Manus—a newer AI research firm being acquired by Meta—as the only other products with what he calls "a true, generally capable, long-running and really smart agent architecture." He argues most other so-called agents are merely "LLM powered workflows," consisting of predetermined steps with AI assistance, not genuine autonomous agents capable of course correction. This is a nuanced claim in a market flooded with new AI agent announcements, from OpenAI's recent agent-building tools to offerings from Notion, Harvey, and countless others. In this crowded field, Liu's assertion that Superagent is fundamentally different must be proven in practice.

In a blog post announcing the product, Liu offers concrete examples. Request an evaluation of Google as a three-year investment, and Superagent provides a structured assessment citing earnings calls, a defensibility analysis against OpenAI and Anthropic, and overlooked risk factors. Ask for a briefing on Wells Fargo's AI strategy before a sales pitch, and it delivers details on regulatory posture, recent AI investments, and specific pain points your product could address. The system leverages premium data sources like FactSet, Crunchbase, SEC filings, and earnings transcripts.

This launch culminates a transformation for Airtable, which Liu has been repositioning as an "AI-native platform." Last fall, the company appointed David Azose, former engineering lead for ChatGPT's business products at OpenAI, as its CTO. Simultaneously, it acquired the AI agents startup DeepSky (formerly Gradient), which had raised $40 million. Superagent will operate with semi-independence from Airtable, led by DeepSky's founding team.

Pricing details were still being finalized recently, but the model appears to follow an emerging standard for AI products: approximately $20 per month per user for an entry-level tier, scaling up to around $200 for power users, with generous inference credits included. "We’re not trying to optimize for profit margin right now," Liu says.

Whether Superagent realizes Liu's vision of a trillion-dollar market or becomes an ambitious bet that fails remains uncertain. The competition is formidable, and the technical distinctions Liu emphasizes between "real agents" and others may not resonate with customers if competitors deliver satisfactory results more quickly and cheaply. Yet, for a CEO whose company has seen a $7.7 billion paper valuation decline while preserving most of its actual capital, this move demonstrates a commitment to investing in the future rather than merely protecting the present.

In fact, Liu has reframed the valuation drop as a recruiting tool, telling employees they are receiving "equity that’s actually much more attractively priced than the $11 billion valuation," with substantial potential upside if his strategies succeed. He has capital for strategic acquisitions and no immediate need to raise more funds. When asked if Superagent represents the larger long-term opportunity, Liu offers a measured response, not dismissing the possibility. Airtable "will probably be larger for at least the near term than any new products that we do, including Superagent," he says. "But I also like being able to bet on Superagent. Optionality is a good thing."

This approach reflects what Liu terms "wartime" leadership—a phrase he once avoided as overly aggressive but now finds fitting. "Being very fast on the draw to be able to adapt," he notes, is "the most value-creative way to run things right now." He quickly adds, "It’s also the most exciting way to do things."