Advertising and media firms today outpace every other industry in generative AI returns, according to a global research report released by Snowflake on March 10, 2026, based on a survey fielded between August 13 and September 17, 2025. The "ROI of Gen AI and Agents 2026" study, conducted by Omdia by Informa TechTarget, covered 2,050 enterprise professionals across nine countries and six industry verticals. For the marketing community, the numbers tell a striking story: ad and media companies are monetizing AI faster than peers in financial services, retail, healthcare, manufacturing, and technology.

The survey methodology is worth noting. Researchers initially contacted 3,479 organizations worldwide and qualified 2,050 respondents who had generative AI solutions in active production. All participating organizations employed 500 or more workers. About 34% represented companies with more than 5,000 employees, 49% came from firms with 1,000 to 4,999 employees, and 17% from organizations with 500 to 999 staff. The margin of error sits at plus or minus 2 percentage points at the 95% confidence level.

The advertising and media advantage

The headline figure is hard to ignore. According to the report, advertising and media organizations report a 69% ROI on their generative AI investments - meaning $1.69 earned for every $1 spent. That compares with a 49% cross-industry average, or $1.49 per dollar invested. Technology companies, often assumed to lead, come in at 58%. Retail follows at 46%, financial services at 44%, healthcare and life sciences at 43%, and manufacturing at 38%.

What explains the gap? The report points to several converging factors. Advertising and media organizations are the most likely of any sector to report deploying generative AI across many use cases - 83% say this, versus a 66% average across all industries. They are more likely to target customer-facing applications (47% versus 36%), meaning their gen AI investments land closer to revenue. Marketing teams in the sector use gen AI at 40% of surveyed organizations versus 35% generally, and software development teams at 53% versus 50%. A particularly striking detail: 83% of ad and media companies apply gen AI to automating code testing processes, compared with 64% across all sectors.

The sector also reports fewer of the foundational challenges that slow other industries. Data quality and quantity problems were cited by just 27% of advertising and media respondents, well below the 40% all-industry average. This suggests these organizations entered the generative AI era with more structured data assets and cleaner pipelines than most.

That said, success creates its own pressures. The report flags that early movers in advertising and media are now bumping into scalability and performance constraints at a higher rate than peers - 35% cite this versus a 27% industrywide average. More pointedly, 55% report that legacy systems are incompatible with modern AI requirements, compared with 38% elsewhere. The same proportion - 55% - cite difficult data preparation tasks like extraction and chunking as an obstacle, versus 40% among other industries. Growth, in other words, is straining infrastructure designed for a different era.

The agentic AI pivot

Beyond raw generative AI returns, the report captures something marketers should watch closely: the acceleration toward agentic AI. Where generative AI responds to prompts, agentic systems can reason, plan, and take action autonomously within defined parameters. According to the report, "agentic AI refers to AI systems capable of independent decision-making and autonomous behavior. These systems can reason, plan and perform actions, adapting in real time to achieve specific goals."

Advertising and media companies are ahead here too. According to the report, 42% of ad and media respondents already have agentic AI in production - above the 31% rate seen in other sectors. And 43% strongly agree that work they currently do will be handled by autonomous agents within the next 12 months, signaling high internal conviction about the pace of change.

Across all industries surveyed, 32% of respondents report agentic AI solutions in production. Senior executives expect a 47% return on agentic investments in the next 12 months, closely mirroring what they achieved in their early generative AI efforts. Organizations with multiple gen AI use cases already in production are significantly more likely to have crossed into agentic territory: 44% of that cohort are using agentic AI versus just 8% of organizations still running initial gen AI pilots. The report concludes that "more tech-forward organizations may open up a sizable lead over competitors" as agentic deployment accelerates.

PPC Land has tracked this trajectory across multiple platforms, including Amazon and Google racing to deploy agentic advertising tools in late 2025. The McKinsey Technology Trends Outlook 2025 separately identified $1.1 billion in equity investment flowing into agentic AI during 2024, with job postings in the space rising 985% from 2023 to 2024.

Enterprise-wide performance numbers

The full dataset - covering all 2,050 respondents across industries - reinforces that generative AI is delivering real, measurable returns, not just anecdotal wins. According to the report, 92% of early adopters say they have seen a positive return on their gen AI investments. Among those who formally quantified returns, the figure is 49% - a 20% increase over the prior year's research, which found a 41% quantified ROI.

C-suite confidence is notably higher than the broader workforce. Fully 75% of C-level respondents from nontechnical business organizations report a positive, quantified ROI, edging out C-level technologists at 70% and markedly above middle managers and senior individual contributors at 53%. Among all C-level leaders, just 5% say returns have been flat, and only a single C-level executive across the entire survey reported quantified negative returns.

Operational efficiency leads as the primary driver of deployment. According to the report, 51% of respondents cited it as a top goal, and 88% of all respondents report material efficiency gains. Innovation outcomes matter almost as much: 44% pursued them, and 83% say they achieved measurable improvement. Customer experience improvements were sought by 40% and reported as achieved by 84%.

The teams most commonly running gen AI in production today are IT operations at 62%, data analytics at 59%, cybersecurity at 53%, software development at 50%, and customer service at 49%. Even in the lowest-penetration areas - sales and procurement - more than 30% of organizations report live gen AI tools. The average respondent says their organization uses gen AI across roughly five business functions.

What is failing - and why

The picture is not uniformly positive. The report is candid about persistent friction. According to the findings, 96% of organizations grapple with significant implementation challenges. The top five are data quality and quantity (40%), employee expertise or skill (35%), integration with existing or legacy systems (31%), scalability and performance(27%), and cost (24%).

Data infrastructure problems cut deep. According to the report, 65% of respondents find it challenging or very challenging to break down AI data silos. A further 62% say the same about measuring and monitoring AI data quality, and an equal 62% about preparing data to be AI-ready. On average, only 20% of organizations' unstructured data assets are considered "AI-ready," and only 32% of their structured data meets that bar. Just 7% of respondents say more than half their unstructured data is AI-ready - a number that actually declined from 11% the year before.

The talent dimension is uneven across company sizes. Midmarket companies with 500 to 999 employees are significantly more likely to cite a skills shortage - 43% compared with 34% of enterprise respondents. The report suggests this gap narrows as companies accumulate AI experience: 40% of organizations with only initial gen AI solutions in production cite employee skills as a challenge, versus 33% of those running many active use cases.

Cost overruns are nearly universal. According to the report, 95% of organizations said at least one component of their gen AI solutions had exceeded budget. Data storage and compute costs were the most common culprit, cited by 60%, followed by supporting software such as logging, monitoring, and developer tools at 55%. Four of the five cost categories measured declined year on year, with one exception: talent costs, which remain elevated and show no sign of moderating.

Organizations are responding by increasing investment. On average, technology budgets allocate about 22% to gen AI projects over the next 12 months - roughly in line with the 23% figure from the prior year's report. Organizations already running many gen AI use cases plan to commit 25% of their technology budget to AI, versus just 16% among those just getting started.

The shadow AI problem

One tension the report surfaces - directly relevant to the marketing industry - is the prevalence of unauthorized AI use. According to the findings, 57% of all respondents acknowledge using non-approved AI tools at work. The rulebreakers are disproportionately senior: 66% of C-level business leaders report using non-approved AI technologies, compared with 55% of tech-side C-level leaders and 56% of middle managers.

The disconnect between business teams and IT functions is particularly stark for marketing. According to the report, 51% of marketers say their teams are using gen AI, while only 37% of IT staff acknowledge the same. In customer support, 63% of those teams report gen AI use versus 51% of IT. The pattern repeats across HR, procurement, manufacturing, and sales.

The risks of unsanctioned use are real. Employees feeding confidential data into external models, violating compliance standards, and receiving unreliable outputs from systems untrained on company data all feature in the report's analysis. Yet the drivers of shadow AI are structural rather than merely behavioral: 60% use non-approved tools because approved alternatives lacked specific capabilities, 53% because approval processes are too slow or frequently rejected, and 37% simply because non-approved tools get the job done faster.

Marketers betting on AI media while struggling with internal AI adoption was a pattern documented by PPC Land in January 2026, drawing on separate Mediaocean survey data showing 42% of marketing professionals face data quality barriers to broader AI implementation - strikingly similar numbers to the Snowflake findings.

Jobs: creation and destruction

The employment picture is nuanced and deserves care in interpretation. According to the report, 42% of respondents say generative AI has only created jobs at their organization. A further 35% say both job creation and job loss have occurred simultaneously, and 11% report only job losses. Thirteen percent say AI-powered automation has not affected employment at all.

The teams most affected by job losses are IT operations (40%), customer service and support (37%), and data analytics (37%). Entry-level roles bear the heaviest burden: 63% of organizations reporting any job loss saw it at entry level, with middle management next at 46%. Senior management saw job losses at just 27% of surveyed organizations.

However, adoption depth matters here too. Seventy-five percent of organizations running many gen AI use cases say the net employment impact has been positive, versus 56% among those in early stages. Among the 35% who report both job creation and loss, 69% characterize the net effect as positive. Job gains are most common in cybersecurity, IT operations, and software development.

Snowflake cofounder Benoit Dageville, writing in the Snowflake Data + AI Predictions 2026 report, raised a long-term structural question about the pipeline of technical talent: "AI has arrived so quickly that we don't yet know how the world is going to rebalance or reorganize itself."

Geographic variation

The report breaks findings down by country, revealing substantial variation in both adoption rates and returns. Germany leads on deployment breadth, with 50% of German respondents saying their organizations use gen AI across many use cases, versus a 39% worldwide average. India leads on ROI measurement confidence at 71%, and reports standout outcomes including 94% seeing improved efficiency and 92% improved innovation versus 87% and 83% globally.

France presents the starkest contrast. French organizations report a 32% quantified ROI - the lowest of any country surveyed - meaning $1.32 earned per dollar invested against a global average of $1.49. Adoption in software development (31% versus 51% globally), customer service (39% versus 49%), and IT operations (47% versus 63%) all lag significantly. Japan presents a different pattern: only 48% of Japanese respondents have formally quantified positive ROI, but among those who have, the figure is 59% - higher than the global average.

For the UK, ethical and regulatory concerns are a primary drag. According to the report, 35% of UK respondents cite these concerns as a key challenge versus 23% globally, which helps explain the lowest gen AI penetration rate recorded across the nine surveyed markets. Yet when gen AI is deployed, UK organizations see a 61% quantified ROI - above the global average.

The agentic infrastructure challenge

For the marketing community specifically, the gap between intention and infrastructure emerges as the central strategic issue. The report makes clear that the primary blockers are not the AI technologies themselves but rather data readiness and talent. According to the findings, respondents "consistently indicated that the challenges they faced were not with the gen AI and agentic technologies themselves. Rather, two core areas are presenting most of the obstacles: talent and data infrastructure."

For advertising and media firms, this translates directly. Legacy systems incompatible with modern AI requirements (55%), difficult data preparation (55%), and interoperability issues (42% across all sectors) create compounding friction as organizations attempt to move from generative AI toward more autonomous agentic deployments. The Newton Research integration with Snowflake Cortex AI announced in November 2025 illustrated one architectural response - enabling brands to run media mix modeling and incrementality analysis directly within secure data environments without data transfer requirements. LiveRamp's agentic orchestration similarly addressed data governance concerns by giving AI agents governed access to identity resolution and audience activation pipelines.

Among organizations currently using agentic AI, the improvements already being reported are substantial. According to the Snowflake report, more than 80% of respondents at agentic-enabled organizations are seeing improvements in creative workflows (89%), customer experience (89%), strategic decisioning (88%), forecast accuracy and scenario planning (86%), and knowledge sharing (82%). These are not marginal gains. They suggest a compounding effect as data infrastructure matures and agent capabilities become more deeply embedded in operational workflows.

The IAB Tech Lab's Agentic RTB Framework released for public comment in November 2025, and the launch of the Ad Context Protocol on October 15, 2025 - built on Anthropic's Model Context Protocol - indicate the industry is building technical scaffolding to support this shift. Whether those standards gain traction, or fragment into competing implementations, will significantly shape how quickly advertising and media firms can convert their current ROI lead into a durable structural advantage.

Timeline

Summary

Who: Snowflake (NYSE: SNOW), in partnership with Omdia by Informa TechTarget as the research firm. The survey covers 2,050 enterprise professionals across organizations with 500 or more employees in nine global markets.

What: A global research report titled "The ROI of Gen AI and Agents 2026" documenting that advertising and media companies achieve the highest generative AI ROI of any industry surveyed at 69%, compared to a 49% cross-industry average, while also leading on agentic AI adoption at 42% in production versus 31% in other sectors. The broader survey finds 92% of early adopters report positive returns, with a cross-industry quantified ROI of 49% - up 20% from the prior year.

When: The survey was fielded between August 13 and September 17, 2025. The report was published on March 10, 2026.

Where: Nine global markets are covered: the United States (41% of respondents), Canada, the United Kingdom, France, Germany, Australia and New Zealand, Japan, Singapore, and India (7% each). Industries covered are advertising and media, retail, financial services, healthcare and life sciences, manufacturing, and technology.

Why: The findings matter because they quantify a performance gap between the advertising and media sector and other industries on AI adoption and returns, while simultaneously documenting the infrastructure and talent constraints that threaten to slow momentum. As agentic AI deployments multiply - with 43% of ad and media respondents expecting autonomous agents to take on current tasks within 12 months - data infrastructure readiness becomes a primary competitive variable for marketing organizations planning their next phase of AI investment.

Share this article
The link has been copied!