StrictlyVC Download Airtable is betting big on a standalone ai agent that could replace its own product
Jan 27, 2026
Howie Liu, co-founder and CEO of Airtable, explains his bold launch of SuperAgent and why autonomous agents matter. He talks about agent architecture, model tuning, real-world use cases like research automation, and why SuperAgent was built as a separate product with integrations and privacy in mind. He also discusses go-to-market thinking and how this bet could reshape Airtable’s future.
AI Snips
Chapters
Transcript
Episode notes
Agents Replace Walls Of Text
- SuperAgent reframes AI from chat outputs to autonomous agents that perform end-to-end work and course-correct.
- Howie Liu argues agents + rich interfaces produce more useful work artifacts than text-only chatbots.
Benchmark And Tune For Domains
- Define domain-specific rubrics and evals to benchmark agent performance and improve iteratively.
- Combine best-of-breed models and tune prompts/skills per domain to close the final-mile gap.
Agent Teams Mirror Expert Firms
- SuperAgent can produce deep, expert-level research by assembling specialized agent teams for multi-step tasks.
- Howie compares that output to hiring a McKinsey but faster and far cheaper.