Wondery unearths profitable insights in troves of podcasts

Love a good podcast? You’re not alone. In 2025, more than 584 million podcast enthusiasts worldwide tune in regularly for their latest episodes. Many are fans of shows such as “New Heights with Travis and Jason Kelce,” “Ghost Story,” or “Flesh and Code,” just three of the 220 shows Wondery, the podcast production house of Amazon, publishes. Meeting audience demand for its popular podcasts requires smart, highly efficient operations, which Wondery achieves with the help of Airtable.

The challenge: Scaling content production, optimizing advertising

Imagine creating metadata for the many podcasts releasing weekly episodes – and writing all the titling, long and short descriptions for different platforms, social content, and even merchandising. “It’s an immense amount of data and manual, time-consuming work,” says Neel Ketkar, chief product officer at Wondery. “It’s challenging for producers, who have lots of other demands on their time.” 

Time is valuable in the fast-paced world of podcast production, where content is consumed the moment it’s released. Producers need more time to deliver more of the content audiences want – and audiences need ways to quickly find it. Additionally, advertisers need data that helps them decide which podcasts to sponsor. These challenges prompted Ketkar to dig in and experiment with Airtable to see what was possible with its capabilities. 

At Wondery, Airtable provides a central hub for the company’s production database, principal production calendar, and schedules. Focusing first on specific use cases rather than end-to-end workflows, Ketkar decided to apply AI-powered Airtable to address four discrete challenges: podcast metadata generation; timestamped chapter creation; internal weekly email reports; and tracking of organic brand mentions. 

Airtable delivered, with built-in AI enabling Wondery to significantly reduce metadata and email preparation, make content more discoverable by audiences, and identify new revenue opportunities for sales.

The solution: Lightening workloads, surfacing ad sales opportunities 

Since the introduction of ChatGPT, Wondery staff has been experimenting with AI. Early explorations included creative ideation for shows and episodes, generating full podcasts with AI, and AI-animating podcasts for YouTube. And while those didn’t stick, the company nurtured a culture that encouraged non-technical people to play with AI tools and explore use cases. So it was no surprise when Wondery’s chief content officer Marshall Lewy used Airtable’s no-code Omni AI feature to create a “metadata generator” app in Airtable.

What was surprising is how quickly Lewy built the app and its impact. “In just 45 minutes, Lewy built the initial app, and I spent 30 more extending it,” says Ketkar. “The AI was an immediate gamechanger: It instantly produces all the metadata we’re looking for.” 

The AI-powered generator ingests podcast transcripts and spits out 20-30 different metadata fields for podcast apps, YouTube, social platforms, and merchandising. It also automatically creates podcast chapters with timestamps. This AI approach to metadata production, tagging, and time stamping not only saves considerable time, but it also improves the quality and search-engine optimization of the output, making content more easily discoverable by interested audiences. 

The Wondery content team also automated developing weekly emails detailing new episode releases with AI in Airtable. That saved an additional 4-5 hours per week to put toward content production. But perhaps Wondery’s most unique application was to use AI-powered Field Agents to assist sales in finding the right advertisers for podcasts. 

To surface potential sales opportunities, the Wondery team used Omni to build a system to identify organic brand mentions. Previously, staff weren’t always aware of all of these – and it would have been incredibly time-consuming to find them. Now Airtable Field Agents review podcast transcripts to catch mentions, analyze sentiment, extract quotes in context, identify the speaker, and organize all the information. This structured, accessible data makes it easier for Wondery’s salespeople to tailor pitches to potential advertisers. Says Ketkar, “With the insights AI agents extract, sales can invite advertisers to reach a target audience who is already engaging with their brand on our platform.”

Advancing the use of AI at Wondery

The Wondery team plans to do more with AI using their existing data in Airtable. “Our central production database, which houses information on hundreds of shows, is a key asset,” says Ketkar. A persistent problem, however, is that, instead of linking to this central database, people copy and paste shows into new, separate databases. By using Airtable’s “views” to access the central database, says Ketkar, the team can experiment more productively with AI. “Using this database, along with specific instructions, we’ll be able to guide the AI in spinning off the correct metadata for other potential applications.”

Ketkar will also try stitching together the smaller, valuable components Wondery has built to create more fully automated end-to-end processes. One might be a workflow where a producer approves a show title and other key information, and then AI automatically generates metadata and seamlessly flows the title into distribution systems. Another might be email automation where AI creates personalized email for listeners, advertisers, or internal teams based on specific data or events captured in the central database.

Adopting AI by starting with focused, valuable solutions and building on them to achieve broader business impact is working for Wondery. “We want everyone to think about how they can use AI in their daily workflows to reduce the time spent on manual tasks and shift to higher value activities,” says Ketkar.