CANNES – Just as brands want to have a singular view of their users and prospects, forward-thinking media investment companies know they can no longer maintain channel-specific teams. It takes big investments in technology—some of it artificial intelligence-enhanced—to accomplish both.
In the past 12 months alone, GroupM launched its global [m]PLATFORM and then bolstered its Xaxis programmatic digital media platform with the machine learning initiative called Co-Pilot and finally the acquisition of Triad Retail Media.
[m]PLATFORM is aimed at “streamlining and simplifying” so that GroupM’s clients can “see their consumers and prospective in the same way across the globe,” Nicolle Pangis, the Global COO of [m]PLATFORM, says in this interview with Beet.TV.
From an organizational standpoint, [m]PLATFORM has brought together teams that used to plan, activate, execute and optimized clients’ investments across channels. This process had historically been done “in different pockets in the agency,” Pangis says. “Now we’re at the point of maturation of our industry that it makes sense to bring those all back together and create a single view.”
The biggest single acquisition by Xaxis was that of Triad, which helps retailers sell ads on their websites, for a sum in the neighborhood of $300 million, as The Wall Street Journal reports. Scaling through Xaxis’s many international offices affords Triad the opportunity to help retailers that don’t already sell ads on their websites.
Co-Pilot represents the entry point to machine learning, which is one aspect of AI. In the simplest terms, Co-Pilot’s contribution to brands engaging effectively with consumers is its ability to predict what may happen when ads are served “or what has the highest probability of happening” based on what it’s seen in the past.
“The great thing about technologies like this is that it can take both data that is specifically advertiser driven or just macro economic driven,” says Pangis.
Like other AI-inspired technology, Co-Pilot “learns as it goes.”
Among its abilities are predicting whether a particular ad impression might be viewable based on history and whether a specific consumer has engaged with ads before.
“As a technology it’s not one that is stagnant,” Pangis says. “It’s not like you put a model in and you leave it forever. You’re constantly putting in models and they’re constantly learning.”