IPG Engine's Next Big Release: An AI Sandbox

As part of Thursday's unveiling of "IPG Engine" -- tied to a new, first-to-market deal between the holding company and long-time partner Adobe -- Chief Solutions Officer Jayna Kothary and Acxiom General Manager of Cloud and Platform Services Jason Brown briefed "Media 3.0" on the engine's roadmap, including a soon-to-launch "sandbox" for rapid development, testing and deployment of new AI tools and applications, as well as early steps they are taking that could finally lead to the ad industry's long-pursued Holy Grail: an "iterative media mix" aligning audiences, data, creative and content assets with media.

While the term sandbox may have garnered some negative industry connotations recently following the IAB's public bashing of Google's privacy sandbox, the term simply means a secure, isolated testing environment where engineers can experiment with and perfect tools and applications before putting them in the wild.



"We've built [the] sandbox to democratize the use of AI across our organization, so people can be better at their jobs and be able to utilize the plethora of tools," Kothary teased when I asked her to describe IPG Engine's roadmap.

She said the sandbox will be formally unveiled soon and that as part of it, IPG will announce new partnerships with other AI technology suppliers, but it will also be used to test proprietary client technology, as well as homegrown IPG tools.

"The way we see the [IPG Engine] living and breathing, is there will constantly be new features and functions and releases. And that will be stuff we’ve built, or partner technology," she explained, noting that the focus of Thursday's announcement with Adobe's GenStudio was on innovating the use of AI in content, commerce and CRM, but future iterations will include new applications of AI for media -- especially integrating media with the rest of client marketing objectives.

In terms of the first-to-market aspects of this week's Adobe announcement, Kothary said there are two new features:

One is the ability to build "custom" AI models for each brand in an IPG client's portfolio.

The other is first holding company access to Adobe APIs (application protocol interface) enabling IPG to engineer new leverage pre-existing data integrations between Adobe and Acxiom, including individual-level identity graph data enabling IPG and its clients to use individual audience IDs to determine who a consumer is, what they are in market for, what offers and creative assets they likely will respond to, and then matching it with the audience IDs of publishers in the wild.

Kothary said IPG will utilize its new AI sandbox in similar ways to experiment with other AI partners, as well as what it develops in-house.

She said those homegrown features will be architected by her corporate team, but that the engineering will be build by IPG Mediabrands' Kinesso unit.

"Our vision is that this steps into a broader, truly omnichannel [solution]," Acxiom's Brown explained, adding, "That word has been thrown around for the last two decades, but this really is a tangible step in that direction of brands being able to go from core first-party data all the way through high-value third-party data to maybe more commoditized third-party data and crafting the optimal mix."

Or as, Kothary explains it, the next step will be to engineer a more seamless integration between client's media plans and their "content matrix."

"There's the potential for an iterative media mix happening, because everything is sitting in the engine," she explained.

Next story loading loading..