Is generative Ai used by advertising agencies?
This fact-check may be outdated. Consider refreshing it to get the most current information.
Executive summary
Yes — generative AI is widely used inside advertising agencies: Forrester data cited by Marketing Dive reports 91% of U.S. agencies are either using (61%) or exploring (30%) generative AI, with 74% naming ideation as a high/critical priority [1]. Industry trade and academic reports show agencies deploy GenAI across ideation, asset production, personalization and programmatic optimization while grappling with legal, ethical and quality limits [2] [3] [4].
1. Adoption is near-ubiquitous, but “use” covers a spectrum
Major industry surveys show near-universal attention: Forrester’s joint survey with the 4A’s found 91% of U.S. ad agencies are using or exploring generative AI, and 61% report active use — but that figure spans light experimentation to deep integration [1]. Trade pieces and vendor posts repeat the headline and stress that “using AI” can mean anything from weekly prompt-based brainstorming to building proprietary, production-grade pipelines [5] [6].
2. Agencies are applying GenAI to both creativity and scale
Agencies report high priority for ideation and creative brainstorming (74% called that high/critical), while nearly half prioritize generating assets for dynamic creative optimization, showing GenAI is used for both conceptual work and high-volume variant production [1]. Case-study lists and agency rankings show practical uses: copy tools for headline variants, image generators for rapid visuals, and automated versioning to hit many channels fast [7] [8] [9].
3. Business model disruption and client pressure are real themes
Analysts warn GenAI threatens traditional billable-hours models by reducing time required for content production; a Boston Consulting Group stat cited in Forbes claims over 70% of CMOs use generative AI for content and personalization, which in turn pressures agencies’ fee structures and value propositions [2]. Agency commentary underscores a counterargument: agencies can retain value through cultural insight, strategic framing and proprietary systems that wrap AI into defensible offerings [2] [5].
4. Agencies are building new capabilities and partnerships
Industry groups and vendor-driven lists highlight investment and partnership activity: the 4A’s produced a dedicated 2025 report on GenAI inside U.S. marketing agencies, and many agencies are forming alliances or stacking tools (commercial or in-house) to scale AI safely and productively [10] [6]. Trade pieces emphasize agencies must convert pilots into repeatable workflows and IP to avoid commoditization [5].
5. Limits: copyrights, hallucinations and human verification required
Academic and sector research documents consistent constraints: copyright uncertainty and ethical concerns constrain adoption; generative outputs require human verification because of factual errors and “hallucinations” [3] [11]. Scholarly work frames GenAI as a complement to creative processes rather than a full replacement, stressing human oversight and co-creative models [3] [4].
6. Emerging tech (agentic AI) promises more change — cautiously
Thought leadership from agency bodies flags agentic AI (multi-agent systems) as the next wave that could reshape workflows and guardrail-setting roles for human teams, but current systems suffer reliability issues and are early-stage, so adoption is prospective rather than widespread today [11]. Digiday and 4A’s commentary indicate firms are experimenting but warn about unpredictability and integration work [12] [11].
7. Competing narratives: threat vs. opportunity
Two consistent narratives appear in the sources: one frames GenAI as a cost- and time-saving disruptor that can erode traditional agency revenue lines [2], the other frames it as an amplifier that forces agencies to reorient toward strategy, ideas and AI-governance to remain indispensable [5] [13]. Both narratives derive from the same data but emphasize either short-term displacement or long-term transformation [2] [5].
8. What reporting does not settle — and what to watch next
Available sources document prevalence of use and typical applications, but do not provide granular, independent measures of how many agencies have fully integrated AI into billing, staffing or IP models; nor do they quantify client outcomes across the market beyond selective case studies (not found in current reporting). Watch metrics in future 4A’s/Forrester follow-ups and peer-reviewed outcome studies for hard ROI, displacement rates and legal-resolution trends [10] [1] [4].
Summary: Agencies are actively using generative AI across ideation, asset production and optimization (91% using/exploring; 74% prioritize ideation), but adoption ranges from experiments to deep integration and is constrained by legal, quality and ethical limits that make human oversight mandatory [1] [3] [4].