Google Ad Manager (GAM) wants to make it easier for publisher ad ops teams to troubleshoot campaigns. And – you’ll never believe it – the solution is an AI chatbot.
GAM is currently beta testing the new customer service chatbot, called Ask Ad Manager, with select publishers, the company announced Thursday. The tool will gradually roll out to more publishers in the coming months, and GAM expects it will be widely available later this year.
At launch, Ask Ad Manger focuses on three problem areas for ad ops. It offers instant troubleshooting help when a campaign isn’t delivering as expected, ideally by diagnosing the problem and suggesting how to fix it. The agent also lets publishers ask more complicated questions about specific bidder performance and how campaigns are running compared to industry benchmarks. And it provides a smart navigation feature that directs publishers to wherever in GAM’s user interface they can make suggested changes or find relevant data.
Ask Ad Manager doesn’t make autonomous decisions; it still requires a human to implement any suggestions. But Google expects the chatbot to evolve alongside other agentic AI tools for buying and selling ads, eventually allowing for publishers to program regular reporting updates and automate some aspects of troubleshooting, said Peentoo Patel, senior director of product management for Google Ad Manager.
New features
GAM already had robust reporting and troubleshooting features for publishers, Patel said. And it has no plans to deprecate any features following the launch of the new chatbot.
But using those existing tools requires more manual work, including managing endless email exchanges and navigating ugly charts and spreadsheets. Querying a chatbot is far simpler, Patel said. Plus, he added, every ad ops team knows the pain of wasting time daily on such tedious processes.
“Every day, someone’s troubleshooting something, or they’re asking for reports and analytics,” Patel said. The goal is to reduce hours spent “doing the back and forth.”
Ask Ad Manger is built on Google’s Gemini AI tech. The chatbot is grounded on the first-party data sourced from the publisher using the tool, as well as generalized benchmarking data that has long been proved by GAM, Patel said. It does not pull in data from other publishers or third-party sources.
In addition, Google said it will honor publisher controls regarding what monetization features can be controlled via GAM, he said. For instance, some publishers might have GAM set up to only handle campaign reporting or billing; in these cases, Ask Ad Manager will not address anything beyond the scope of those approved tasks.
How it works
Using the tool is pretty straightforward.
Say a publisher ad ops team notices a particular line item on a campaign is producing lower yield than expected or isn’t delivering at all. A team member can plug that campaign line item number into the chatbot and ask it to diagnose the cause. The chatbot should then surface whether the problem is as simple as a piece of ad creative failing to render or whether there’s something more complicated going on behind the scenes.
Previously, addressing such a problem could be a multi-hour or even multi-day process, Patel said – and all the while, that line item would be failing to deliver for the publisher.
And the Ask Ad Manager bot can do more complicated problem-solving when it comes to setting pricing floors or creating new programmatic deals, he said.
For example, an ad ops team can query the tool on how a particular bidder is performing compared to other bidders, and whether pricing adjustments would increase that bidder’s win rate. It can help publishers quickly identify which part of their inventory or audiences should be included in certain deals.
Then, once the chatbot has made its suggestions, the navigation feature tells ad ops teams how to make the changes. The platform did not previously offer any similar guidance directly inside its user interface, Patel said.
The end result is more time freed up to spend on yield optimization tasks, he said, and less time troubleshooting campaigns that underperform or go awry.
Costs and concerns
Ask Ad Manager is currently free to use. And Google isn’t limiting how often publishers can query the chatbot during the beta test. But, according to Patel, that’s likely to change as the tool becomes widely available.
Google picked which publishers are participating in the beta test, as is typical of GAM’s publisher testing program, Patel said. And the beta-testing pool includes a mix of large and small publishers across desktop web, mobile and CTV. Yahoo is the only beta tester GAM has named so far.
The testing phase began in mid-June. Because it’s only been running for a couple of weeks, Patel said GAM could not provide any data on benefits the testers have seen from the tool so far.
He also said it’s too soon to discuss how often the chatbot makes unhelpful suggestions or possibly hallucinates data in its responses. But, he added, GAM is receiving feedback from the beta testers and will keep an eye out for these issues as the tool enters the wider marketplace.
But the main thing GAM will be watching is whether the chatbot helps improve publisher yield.
“Publishers are always asking us for better tools for troubleshooting, better ways to navigate our UI,” Patel said, “but also ways for them to make more revenue.”
