Bing Ads is building a bot to help you manage campaigns better

There are now third-party bots available in Slack, Facebook Messenger, Google Sheets and elsewhere that marketers can use to get information about campaign performance and, in some cases, take action.

At the Bing Ads Next event at the Microsoft campus in Redmond, Wash., a few weeks ago, executives spoke extensively about Microsoft’s work in machine learning, artificial intelligence and bots. So what about a bot for Bing Ads? I posed that question in an interview with David Pann, general manager of the search network at Microsoft.

They’re working on it.

Currently dubbed “Associate Account Manager in the Box,” the idea is to give every Bing Ads advertiser access to a virtual account manager.

“Through our bots and artificial intelligence and our machine learning, and through just standard messaging, every customer should get an account manager,” said Pann. “It won’t just tell you how you’re doing, but can make recommendations on what you should be doing to get better performance. We’re starting to build that.”

Pann says they’re starting with an overall account management and health check-in. “Then you can imagine from there automating a lot of the back office and billing capabilities.”

Beyond Associate Account Manager in the Box, Microsoft, not surprisingly, already sees AI and machine learning playing a bigger and bigger role in campaign management. For example, Power BI integrations with Bing Ads surface answers to complex questions internally on behalf of advertisers, and dynamically generated ads pilot for the intent network Microsoft is building out for native ads.

“I don’t have to be the person who owns ‘fly to Boston cheap’ versus ‘cheap flights to Boston”‘

Pann has seen a significant shift in the way marketers have come to see automation. During an introduction to a new testing and optimization product called Decision Service, Pann asked the group of search customers if they would let an algorithm change their creative. The answer was a resounding yes.

“A couple of years ago, people would say no,” said Pann. “But an algorithm can do A/B testing on thousands of permutations so much faster than individuals. And the problem is not those individuals, it’s not that they’re not capable, it’s just that they don’t have time … I think people are realizing there’s so much to do and some of this is so labor intensive. You know, I no longer have to be the person who owns ‘fly to Boston cheap’ versus ‘cheap flights to Boston’ and the thousands of permutations that can occur.

“In search, we’ve just been associated with these things for so many years it’s a natural evolution for us to say the algorithm has been predicting what ad to show on the page, why not have an algorithm create the ad we show on the page and then change it at 1:01, 1:02, 1:03 p.m., or dynamically change it based on the type of user that’s going to view it. Do I think we’re going to get there? Yeah, I do.”