Ad Spend Is Up 8.6%. Agency Revenue Is Down. Here Is Where the Money Went.
Ad spending grew 8.6% while holding company revenues fell 1.2%. Clients are redirecting budgets to AI tools. Most agencies responding wrong.
Global ad spending grew 8.6% year over year in 2025. eMarketer
Over the same period, holding company revenues fell 1.2%. eMarketer
The money did not disappear. It moved. And it moved in a direction that most agencies still have not fully accounted for: directly into AI tools that do work agencies used to own.
industry-specific AI automation__
The Budget Redirect
Three data points frame the scale of the shift.
First: 60% of senior US marketing leaders say they spent less on agencies in 2025 specifically because of AI. Not because budgets shrank. Because they found faster, cheaper alternatives to the services agencies were selling. Typeface
Second: 83% of marketing leaders say they would further reduce agency spending if they could fully automate content creation. 11% would stop using agencies entirely. eMarketer
Third: 73% of teams that have adopted AI agents have already cut their content creation spending on agencies. Not "would cut." Have cut. eMarketer
Marketing budgets themselves dropped from 11% of company revenues in 2020 to 7.7% in 2024. Of that shrinking slice, 9% now goes directly to AI tools, the fastest-growing category in marketing spend. eMarketer
Agencies are being squeezed from both directions. Total marketing budgets are smaller. The portion that used to flow to agency retainers is being redirected to platforms that do the work in-house.
The Consolidation Response Is Not a Fix
The industry's first response has been consolidation. The Omnicom-IPG merger in late 2025 created the world's largest holding company. An estimated 4,000 positions were eliminated with $750 million in expected cost savings within two years. J.P. Morgan
Omnicom, WPP, Publicis, and others have committed hundreds of millions to AI development. eMarketer
This works for holding companies with the capital to build proprietary AI infrastructure. It does not work for the thousands of 10-to-100-person agencies that make up the majority of the industry. Those agencies cannot consolidate their way out of the squeeze. They need to operate differently.
Why Buying More AI Tools Is Not the Answer
The natural response for mid-size agencies is to adopt AI tools: content generators, automated media buyers, AI-powered analytics platforms. Buy the subscription. Roll it out. Hope it keeps the lights on.
The results are not encouraging. MIT found that 95% of enterprise generative AI pilots fail to deliver measurable profit impact. Fortune
RAND's analysis puts the overall AI project failure rate at 80.3%, with 33.8% of projects abandoned outright, 28.4% delivering no value, and 18.1% unable to justify their costs. Pertama Partners
Agencies are not immune to these failure rates. They face the same root causes that sink enterprise AI projects: fragmented data spread across CRM systems, creative asset libraries, campaign platforms, and analytics dashboards. No single AI tool can bridge those silos. 44.4% of marketers cannot find people with the skills to make AI work in their workflows. Campaign US
The pattern repeats across industries. Real estate firms run five AI pilots simultaneously and achieve transformation in 1% of cases. Professional services firms buy AI tools for research and billing but never connect them to their actual workflow. Agencies buy content generators that produce volume without strategy, analytics tools that nobody reads, and automation platforms that sit alongside the existing stack instead of replacing the manual processes.
The tool is not the problem. The integration is.
The Agency Survival Equation
The agencies that are growing through this shift share a common characteristic. They stopped positioning as production houses competing on output and started positioning as the team that helps clients integrate AI into their own operations. eMarketer
This is not a theoretical shift. The data demands it. If 73% of teams with AI agents have already cut agency content budgets, agencies selling content production are fighting a structural decline. If 83% of marketing leaders would cut further with full automation, selling speed and volume is a race to the bottom against software.
The agencies capturing new revenue are the ones that help clients with what AI tools cannot do on their own: data governance, workflow integration, quality control, strategic alignment, and managing the transition from manual to automated operations. These are the capabilities that MIT identified as the actual failure points. 73% of failed AI projects lack clear executive alignment on success metrics. 68% underinvest in data governance. 56% lose C-suite sponsorship within six months. Fortune
Agencies understand their clients' marketing workflows better than any external AI consultant. That knowledge is the competitive advantage, but only if the agency can operationalize it through AI systems rather than through headcount.
What Proper AI Integration Looks Like for Agencies
91% of senior agency leaders expect AI to reduce headcounts. 57% have already slowed or paused entry-level hiring. eMarketer
Hiring freezes are a cost response, not a capability response. They reduce burn without building anything new. The agencies that survive the next two years will be the ones that redirect that headcount savings into AI systems that do three things:
First, connect the fragmented data. Campaign data, client CRM records, creative performance metrics, and audience insights need to flow through a single analytical layer. Not a dashboard. An operational layer that makes decisions and routes work based on what the data says. This is the implementation sequence that most organizations get wrong.
Second, automate the repeatable operations. Media reporting, campaign setup, A/B test analysis, content repurposing, client status updates. The work that consumes junior hours without requiring senior judgment. Automating these tasks does not eliminate value. It redirects human time toward the work clients actually pay premium rates for: strategy, creative direction, and business growth.
Third, package AI integration as a new service line. If clients are pulling budget to build their own AI capabilities, agencies should be the ones building it for them. Marketing operations, content systems, analytics infrastructure. The agency already understands the client's business. Adding AI implementation to the service portfolio captures the budget that is currently flowing to generic SaaS tools or big-four consulting firms charging significantly more for the same work.
The Window Is Specific
Canadian small business AI adoption doubled from 6.1% to 12.2% in a single year. Harris Poll That number will double again. The SMBs making adoption decisions right now will choose their partners in the next 6 to 12 months. Agencies that can offer AI integration alongside creative services will capture those partnerships. Agencies that wait will compete on price against tools that get cheaper every quarter.
The assessment process is designed to identify exactly where an agency's existing workflow can be connected to AI systems for operational gain. Not a demo of a tool. An analysis of the specific integration points where automation creates measurable value.
The ad spend is growing. The question is whether your agency is structured to capture where it is going.