Airtable's AI-Native Refounding

This refounding is Airtable trying to move from a tool people configure by hand into a system where work starts in conversation and ends in a live app that can also run the work. Omni turns a plain language request into a base, interface, and analysis flow, while Field Agents do the repetitive research, enrichment, and routing inside that app. That matters because it shifts Airtable closer to an operating layer for knowledge work, not just a prettier spreadsheet database.

• The product change is paired with an infrastructure change. Airtable rebuilt around reusable app components and HyperDB, which supports up to 100M rows and connections to systems like Snowflake, Databricks, and Salesforce. That makes AI generated apps more viable in large enterprises, where toy datasets break quickly.
  1 sacra 5 airtable 6 airtable
• The competitive set is widening. Airtable now overlaps more directly with Zapier on automation, with Notion and Coda on conversational app creation, and with Glean, Writer, and Hebbia on AI assisted knowledge work. The common battle is to own the interface where a user asks for work and the software actually executes it.
  1 sacra 4 airtable 7 sacra 8 sacra
• The business logic is expansion. Airtable was already at an estimated $478M ARR in 2024, with enterprise growth outpacing work management peers and enterprise net dollar retention reaching 170%. Making AI the default interface gives Airtable a cleaner path to sell more seats, more workflows, and eventually standalone AI products like Superagent.
  1 sacra 3 sacra 9 sacra

The next step is Airtable stitching these pieces into a full loop, where Omni builds the app, agents populate and update it, and Superagent style research flows feed decisions back into the system of record. If that loop works, Airtable becomes harder to displace because it will own both the structure of work and the intelligence acting inside it.