Marketing AI Pulse Brief (May 2026): The CMO Scaling Gap, AI Search, Agentic Buyers and the End of Manual Ad Management
In the May 2026 edition of The Marketing AI Pulse Brief, host Aby Varma of Spark Novus and co-host Matt Cyr of Loop AI Agency go deeper on fewer stories — covering the developments that most deserve more than two minutes of airtime. Drawing on new Gartner research from the London Marketing Symposium, Google's announcement of the most significant search overhaul in 25 years, Harvard Business Review's study of AI shopping agents across 16,000 simulated purchasing rounds, and a wave of simultaneous moves by Google, OpenAI, and Meta to automate paid media execution, this episode examines the structural shifts CMOs need to understand and act on now.
CMOs Are Spending on AI but Can't Scale It
Gartner surveyed 402 CMOs and found that AI automation of marketing work is expected to more than double — from 16% today to 36% by 2028. Marketing leaders are already directing 15.3% of their budgets toward AI. The problem: only 30% say they are actually ready to scale it.
Gartner's term for what's happening to the other 70% is "AI competency traps." These are organizations spending to feel prepared rather than to move faster. The CMOs who are pulling ahead — what Gartner calls "market-shapers" — are not spending more. They are spending differently. They have moved from experimentation to execution, and the gap between them and everyone else is accelerating, not holding steady.
Matt Cyr addressed this exact dynamic at the Build a Better Agency Summit in Denver, where his talk carried the working title "Death by a Thousand Cuts: How Disconnected AI Efforts Are Slowly Bleeding Your Agency Dry." The theme resonates beyond agencies. Organizations are innovating and experimenting — but too few are taking the next step from a successful pilot to a scalable system. When experimentation happens without an overarching strategic direction, you end up in an endless cycle. Different tools, different teams, different experiments — many of which succeed individually but never get scaled because other priorities always seem more urgent.
The implication is direct. Spending 15% of your budget on AI while only 30% of teams can scale it means the majority of that investment is producing pilots, not performance. The key number to watch is not how much you are spending — it is whether that 30% scaling rate is going up. If the budget keeps growing while the scaling rate holds flat, the gap between investment and return will become impossible to ignore.
Marketing leaders should run an honest AI competency audit: map every AI investment on a spectrum from experimentation to scaled execution. Prioritize use cases that deliver the most value with the least complexity. Tie every AI line item to a specific workflow, metric, or customer outcome, not a general readiness goal. Then ask the harder question honestly — are you a market-shaper, a market-conformer, or a late adopter? Benchmarking against your own prior year is not enough. Benchmark against category leaders.
The Search Box Is Being Retired
On May 19th, at Google I/O, Google announced the most significant overhaul of Search in 25 years. The traditional search box is being replaced by an AI-powered interface built on Gemini. This was not a roadmap announcement. Rollout has already begun.
The scale of what is already live is worth stating plainly. AI Overviews now appears on 48% of all Google searches, up from 34% just five months ago. AI Mode has 1 billion monthly users. AI Overviews has 2.5 billion. These are not fringe features. They are the mainstream experience. And the numbers that marketers need to internalize: 58.5% of all US Google searches now end without a single click, and the click-through rate for the top organic result has dropped from 27% to as low as 11% on queries where an AI answer appears above it.
Google also announced information agents — AI that runs in the background 24/7, monitors the web, and proactively alerts users when something relevant to them changes, interpreting the significance of that change based on how the user has configured it. Users do not have to search at all. This is coming this summer for US subscribers. Think of it like Google Alerts, but with the comprehension of a well-briefed analyst rather than a keyword matcher.
The traditional search box that marketers have built strategies around for 25 years is becoming what the corded telephone on the wall became to mobile phones — a relic we will eventually look back on and struggle to imagine ever using as a primary interface.
Four days before I/O, on May 15th, Google published their first official AI Optimization Guide. The headline is important: AEO and GEO are not separate disciplines. Google's exact words — "optimizing for generative AI search is still SEO." They also explicitly named what does not work: llms.txt files, chunking content for AI, and generating fake mentions or citations. Google says its systems ignore all of it. It is worth noting that some practitioners are seeing different results in their own experimentation — particularly with structured content and llms.txt — and the picture may vary across platforms like Claude, Perplexity, and ChatGPT. This guidance will require ongoing unpacking as the landscape continues to evolve.
Brands being cited inside AI answers are earning 35% more organic clicks than brands that are not. The question marketers need to be asking has changed. It is no longer "how do we rank on page one?" It is "how do we get cited in the AI answer?" Those are different problems requiring different solutions.
A useful framework for thinking about this comes in three buckets. The first is technical: make sure your site is indexable, not blocking LLM crawlers, and that structured data like schema markup and FAQ sections are in place so AI systems can interpret your content accurately. The second is content: helpful, authoritative, and authentic content remains the foundation. If an AI would not cite it, neither would a discerning reader. The third is PR: in AI search, credibility is measured in part by how many authoritative external sources reference your brand. This mirrors the old backlink logic, but the emphasis is on earned credibility from recognized sources rather than link volume alone.
The practical starting point is measurement. You cannot improve what you do not measure. HubSpot launched a free tool called AEO Sensor at hubspot.com/aeo-sensor that shows how your brand appears in ChatGPT, Gemini, and Perplexity answers right now, with no subscription required. Get a baseline, then build your content and PR strategy around closing the gaps. As Google made clear — it is still SEO. Stay consistent on the fundamentals.Your Customers Are Becoming AI Agents — and Your Persuasion Tactics Don't Work on Them
Harvard Business Review published research this month that should prompt a serious conversation in every marketing leadership team. A growing share of online shoppers are now AI agents — researching, comparing, and purchasing on behalf of consumers. HBR tested eight common promotional tactics across four AI models and more than 16,000 simulated shopping rounds.
Scarcity badges, countdown timers, strikethrough pricing, and bundle framing were all unreliable. Some actively backfired. The only tactic that worked consistently across all models was star ratings. Advanced models like GPT-5 and Gemini 2.5 Pro did not just ignore persuasion tactics — they penalized them, appearing to read psychological pressure signals as indicators of low quality.
The implication is structural. The persuasion playbook that digital marketing has relied on for two decades was built on human psychology — loss aversion, scarcity bias, anchoring effects. AI agents do not have those biases. Each AI model also behaves as its own audience segment, responding differently to the same signals. There is no single "optimize for AI agents" solution. And this raises a genuinely open question: just as human psychology created an entire discipline of behavioral marketing, could we eventually see something like AI psychology emerge — where brands study the behavioral tendencies of specific models and tailor their signals accordingly? It is early, but it is worth watching.
What this means practically: star ratings are a proxy for real human endorsement, and AI agents treat them as such. Competitive pricing and authentic reviews are the two signals that worked consistently across all models — make sure yours are strong. Audit your persuasion stack and identify which tactics are built on manufactured urgency or psychological pressure. Those need to be reconsidered for AI-driven contexts. And build a testing practice. AI model behavior changes with every update. Whoever builds the infrastructure to regularly run agents against their product pages first will have a durable advantage.
Manual Ad Management Is Being Phased Out by Every Major Platform at Once
The scale of what happened in paid media this month is unusual even by recent standards. Google, OpenAI, and Meta all made moves in the same direction — toward AI-managed execution and away from human-managed tactics.
Google announced that Dynamic Search Ads, one of the most widely used campaign types in the industry, are being retired and replaced by AI Max. AI Max is a campaign-level setting where Google's AI takes over keyword matching, writes the ad headlines and descriptions, and adjusts landing page destinations based on intent. The human sets the goal. The AI determines how to get there. Migration of existing campaigns starts in September 2026. This follows a pattern Google has been executing all year — standard text ads, form ads, and responsive image ads were all retired earlier in 2026. Manual ad formats are being removed systematically. Google also announced brand new ad placements that only exist inside AI Mode, including sponsored responses and Direct Offers that surface when purchase intent is detected in an AI conversation.
On May 5th, OpenAI launched a self-serve Ads Manager at ads.openai.com. Any brand can now buy ads directly inside ChatGPT on a cost-per-click model with no minimum spend. Major holding groups including WPP, Publicis, Omnicom, and Dentsu are already integrated. OpenAI's stated goal is $2.5 billion in ad revenue this year.
Meta launched AI Connectors in open beta, allowing advertisers to manage campaigns, audiences, creative, and reporting directly through AI tools like Claude and ChatGPT using Meta's own MCP server, with no developer credentials required.
The skills that have defined paid media for the past decade — building keyword lists, managing bids, writing copy variations, structuring campaigns — are being automated at the platform level. The CEO of one leading AI media buying platform put it plainly when asked what his most successful clients look like: "Turn the AI on and leave it alone." The humans set the strategy, the budget parameters, and the brand goals. After that, the AI is better at the mechanics of buying, optimizing, and finding new markets than any human will be.
The marketer's role is not disappearing, but it is shifting in a very specific direction. The value is moving from execution to inputs. The teams that win will be the ones best at briefing the AI — giving it sharper creative direction, cleaner first-party data, more precise audience definitions, and clearer business goals. Garbage in still means garbage out. Now it gets deployed automatically at scale.
If you are running Dynamic Search Ads today, audit those campaigns now and develop a transition plan before the forced migration in September. Run a test on ChatGPT Ads Manager this quarter — the cost-per-click model with no minimum spend makes it easy to measure, and the barrier to entry is low enough that there is no reason to wait. And have a measured, honest conversation with your paid media team about what their role looks like going forward. A lot of experienced paid media professionals are understandably concerned. The right response is not to move fast on headcount decisions — it is to understand where that expertise can be redeployed. Data quality, creative thinking, audience insight, and business context are all things great paid media people bring to the table. Those capabilities are more valuable now, not less. Be deliberate before drawing conclusions about roles, and let the situation develop further before making structural changes you may regret.
Let's Talk About What This Means for Your Team
The AI shifts covered in this brief are not theoretical — they are happening now, and the gap between organizations moving intentionally and those still in pilot mode is growing every quarter. If you want to think through what any of this means for your marketing strategy, budget, or team structure, we would welcome that conversation. Contact us to set up time with our team.
Frequently Asked Questions
-
Google launched an AI-powered search interface built on Gemini that replaces the traditional text search experience. AI Overviews now appear on 48% of all Google searches, and 58.5% of US Google searches end without a click — meaning users get their answers directly from AI without visiting a website. Brands that are cited inside AI answers earn 35% more organic clicks than those that are not, making visibility in AI-generated responses a core marketing priority.
-
AEO stands for Answer Engine Optimization — the practice of structuring content so AI systems cite it as a source in their generated responses. Google’s own May 2026 guidance states that optimizing for AI search is still SEO. The difference is intent: instead of asking whether content ranks for a keyword, AEO asks whether an AI would cite it as an authoritative answer. Content with original data, first-hand expertise, and a clear point of view performs best.
-
Tactics like scarcity badges, countdown timers, and strikethrough pricing are built on human psychological biases — specifically loss aversion and anchoring. AI shopping agents do not have those biases. Harvard Business Review research across 16,000 simulated purchasing rounds found that advanced AI models actively penalized psychological pressure tactics, interpreting them as signals of low quality. The only two signals that consistently influenced AI agent purchasing decisions were competitive pricing and authentic star ratings.
-
AI Max is a campaign-level setting in Google Ads where Google’s AI manages keyword matching, ad copy generation, and landing page selection based on user intent. It replaces Dynamic Search Ads, which are being retired. Migration of existing campaigns begins in September 2026. AI Max is part of a broader pattern — Google has already retired standard text ads, form ads, and responsive image ads in 2026, systematically removing manual ad formats.
-
Yes. OpenAI launched a self-serve Ads Manager at ads.openai.com in May 2026. Any brand can buy ads directly inside ChatGPT on a cost-per-click model with no minimum spend and no agency required. OpenAI has set a target of $2.5 billion in ad revenue for the year, and major agency holding groups including WPP, Publicis, Omnicom, and Dentsu are already integrated with the platform.
-
Marketing leaders should start with an honest audit of where their AI investments sit on the spectrum from experimentation to scaled execution. Every AI budget line should be tied to a specific workflow, metric, or customer outcome — not a general readiness goal. The organizations pulling ahead are those that have moved from running pilots to running repeatable, measurable AI-powered workflows. Benchmarking against category leaders rather than your own prior year is essential to understanding where you actually stand.
-
As Google, OpenAI, and Meta automate keyword targeting, bid management, and ad copy generation at the platform level, the highest-value contribution for paid media teams shifts to the quality of inputs they provide the AI. That means sharper creative direction, cleaner first-party data, more precise audience definitions, and clearer business objectives. The platforms handle execution. The team’s competitive advantage comes from what they bring to the machine.
