These agents will do everything, if they’re built on Google systems can it tweak them to help Google?

Robert Webster, TAU Marketing Solutions

Despite several tussles with the Justice Department and hand-wringing over the fate of third-party cookies, the company’s stock price popped last week, as the markets gauged Alphabet’s revenue plans following the Google I/O and Google Marketing Live announcements. 

The Google Marketing Live updates include several AI-powered advertising products and capabilities across Search and YouTube (see below). The offerings focus on simplifying creative production, enhancing targeting precision, heralding predictive intelligence, with Google’s Vidya Srinivasan, vp and general manager, ads & commerce, Google, claiming they can “remove the guesswork and drive measurable impact.”

Key product launches and capabilities include:

• Ads in AI Overviews and AI Mode on Search: Google is expanding its AI-powered ad formats to desktop AI Overviews and introducing ads within AI Mode. These integrations aim to align ads more closely with how users naturally explore and discover information, offering marketers new visibility in emerging search behaviors.
• Smart Bidding Exploration: This marks Google’s most significant bidding update in over 10 years. It uses AI to identify and bid on less obvious, high-performing search opportunities, helping advertisers reach consumers in unexpected but impactful moments.
• AI Max for Search campaigns: Originally introduced to enhance campaign performance, AI Max is being expanded to help advertisers better align with shifting user behaviors and unpredictable decision-making journeys.
• Agentic AI capabilities: To lower the barrier to entry for smaller businesses, Google introduced personalized AI agents that guide marketers in optimizing their campaigns, making advanced tools more accessible and manageable.
• Creative generation with Veo and Imagen: Google’s advanced video and image generation models, Veo and Imagen, are being embedded into Google Ads and Merchant Center. This enables advertisers to produce high-quality visual content more efficiently, accelerating campaign development with minimal manual effort.

These product updates reflect Google’s broader strategy to embed AI across the advertising lifecycle — from discovery and creative to delivery and performance — making ads more relevant, actionable, and efficient.

When Google speaks, the entire advertising industry should listen, but as several disclosures in the antitrust trials of the last two years show, this should never be in an obsequious manner. Several Digiday sources attempted to read between the lines of Google’s product announcements last week, and concluded that it is trying to define measurement and control optimization as the AI wars with Big Tech continue. This is the playbook that led to its colossal success, and some fear history may repeat itself.       

Nathan Woodman, founder of Proof-in-Data, observed how the two key innovations, i.e., new AI formats enhancing user experience and measurement tools like AI Max, although he warned the tools may prioritize Google’s revenue growth over advertisers’ success.

“There’s a lot of people who have a problem with handing the keys over to Google,” said Woodman. “In many cases, you get the people inside an organization doing the media buying, they look at the dashboard, they look at the screen… they can lose a critical eye as to what’s happening.”

According to Woodman, who formerly held senior roles at several digital media agencies, this can lead to “a GPS-effect,” a phenomenon whereby “the GPS is telling you where to go, but you don’t really look around.”

Several sources noted that while the potential for the latest announcements to help marketers is clear, but questioning the provenance of these news systems is critical, as “blindly following the GPS is detrimental.” 

Key questions to ask Google about its AI systems

Digiday sources advised marketers to interrogate the below issues when considering the latest Google AI tools. This is because Google’s strategy involves integrating AI agents to control budgets, bids, planning, and creative decisions, and Google has a proven track record for favoring its own systems.

• Analyze the interoperability of Google’s AI agent tools with advertisers’ own data and systems, as well as with non-Google publishers.

• Assess the transparency and potential biases in how Google’s AI agents make recommendations and decisions.
• Evaluate the data privacy and control implications of using Google’s AI agent tools, and how to ensure advertisers’ data is not misused.

Robert Webster, CEO of TAU Marketing Solutions, noted how Microsoft and Google are positioning their tools to control the web, with this competition likely to prove a concern at Google HQ as many doubt its ability to monetize search in the AI-era compared to its earlier fortunes. 

“What everyone is fighting to be is the orchestration layer… where you use their agents to do things,” he added. “Microsoft’s hook is that everyone has a license for Word and Excel, so you’ll use them to do it. Google’s hook is that you spend most of your money with Google, you’ll want to use their agents to connect to Google Ads and your data seamlessly.”

Webster, formerly a senior executive at WPP’s media investment arm, notes that this AI arms race should cause concern among regulators — Bloomberg likewise reported such DOJ scrutiny last week — with the concerns raised about Google’s AI an echo of those documented in the most recent antitrust case. 

“From a DOJ point of view, isn’t that a control game… these agents will do things like decide budgets, decide bids, decide planning, and what creative to run,” observed Webster. “If they’re built on Google systems, will Google be able to tweak them to help Google?”

What we’ve heard 

“Everything is broken, because it’s designed to be that way… There’s a lot of people out there who are just afraid to ask the right questions, and AI is only going to intensify that.”

– A veteran media agency exec discusses how history seems to be repeating itself at the recent Digiday Programmatic Marketing Summit, where agency execs grumbled over measurement and attribution.

Numbers to know 

• 400 million: How many monthly active users Gemini app now has

• €118.9 billion: European digital advertising in 2024
• 20%: Percentage that e-commerce sales are forecast to account of worldwide retail sales, despite slowdown
• 53.6%: Percentage of total global ad spend (excluding China) that is accounted for by Alphabet, Amazon and Meta
• $77 billion: Capex remained elevated during Q1 among Microsoft, Amazon, Google and Meta, driven by AI

What we’ve covered: follow these key stats

WTF just happened to the IAB Europe’s TCF — and what does it mean for targeted ads in the EU?

• Belgium’s Court of Appeal ruled that IAB Europe’s TCF isn’t outright illegal but confirmed that consent strings can be personal data and IAB Europe is jointly responsible for how that data is processed within the framework.
• The ruling maintains pressure on the ad industry to meet GDPR standards, signaling that cosmetic compliance isn’t enough — advertisers must prioritize transparent, permission-based data practices to maintain trust and legality.

What we’re reading

Perplexity is burning through cash, quickly

While the AI-powered search engine has been able to increase its subscription revenue and users in the past 12 months, that growth has come at a large cost. According to The Information, Perplexity generated $34 million in revenue in 2024, but burned through $65 million in cash, which was spent on cloud servers and AI models which power much of its answers.

Meta’s attempt to get its antitrust case thrown out failed

The FTC’s battle against Meta kicked off on April 14, but Meta has since tried to have the case  thrown out. However, on May 20, Judge Boasberg denied Meta’s motion, which asked him to rule against the FTC before it had even launched its defense case, according to The Verge.

As TikTok’s U.S. existence is still in limbo, Shop staff are being laid off

U.S. TikTok Shop staff were told to work from home on May 21, as they waited to receive emails regarding “difficult decisions,” which everyone understood to mean job cuts, according to Bloomberg. The layoffs follow leadership changes, including those in TikTok Shop which saw Mu, a former exec of Douyin’s (TikTok’s Chinese counterpart) e-commerce arm, take over the platform’s U.S. operations.

‘Marching off a cliff’: Developers at Microsoft Build question their future relevance’

Microsoft Build presenters celebrated the company’s latest advancements in AI, including multi-agent AI orchestration in Copilot Studio, the NLWeb protocol for natural language web interaction, and the integration of xAI’s Grok models into Azure AI Foundry.  All these emphasize a shift toward an open agentic web. However, as Semafor’s Rachyl Jones noted, the mood on the floor was not so bullish, noting how one attendee asked: “Is there going to be a Build 2035, or will there not be any more developers?”