DoubleVerify's latest survey of marketing decision-makers finds artificial intelligence spreading through almost every stage of media buying, while the hours spent adjusting campaigns by hand have barely moved. Marketers spent an average of 9 hours and 58 minutes a week on manual optimizations in 2026, according to "Beyond the Hype: The Real State of AI in Digital Advertising," a report DoubleVerify sent to marketers on October 29, 2026. The company's previous survey put the figure at 10 hours and 12 minutes.

The same report finds that half of marketers now use third-party AI to optimize bidding, up from 42% in 2025. It comes from a company with a stake in the answer. DoubleVerify sells its own bidding algorithms under the DV Scibids AI name, and several sections of the report double as product descriptions.

In Short

DoubleVerify, a company that checks where online ads appear and also sells software that sets ad bids automatically, asked about 2,000 marketers how much they use AI to plan and buy ads. Half of them now use outside AI tools to set their bids and almost all use AI somewhere, but they still spend roughly ten hours a week fixing campaigns by hand, about the same as a year ago. That gap matters because saving time is the main promise of these tools, and the company reporting the numbers is also one of the companies selling them.

What DoubleVerify published

The report belongs to DoubleVerify's 2026 Global Insights series. It reached marketers by email on October 29, 2026, under the subject line "New Global Insights report on AI in media buying." The email opens with a question - "Is AI in media buying living up to its promise?" - and says the report "draws from our annual global survey of marketing decision-makers."

Four themes structure the document: adoption and trust, the media planning and buying tasks handed to AI, bidding optimization, and what DoubleVerify calls challenges and concerns. It ends with a description of DoubleVerify's own products and a short case study from Suntory.

This is the second report this month built on the same marketer survey. On October 14, DoubleVerify published a report on social and video platforms, which PPC Land covered under the headline Social purchases drop to 25% from 30%, DoubleVerify surveys show. That edition paired the 2,020 marketers with 22,000 consumers. The AI report uses only the marketer side.

There is also a direct predecessor. On August 13, 2025, PPC Land reported that DoubleVerify's 2025 study found marketers spending 10 hours weekly on manual tasks. Several of the 2026 metrics revisit that report, though not all of them line up, as the sections below show.

How the survey was run

Sapio Research carried out the fieldwork in February 2026, according to DoubleVerify's methodology page, using an email invitation and an online survey with respondents drawn from online partner panels. The sample consists of 2,020 marketing and advertising decision-makers in 25 countries:

  • Asia Pacific: Australia, New Zealand, India, Indonesia, Japan, the Philippines, Singapore, Thailand and Vietnam
  • Europe, Middle East and Africa: Austria, France, Germany, Italy, the Netherlands, Poland, Saudi Arabia, Spain, Switzerland, the United Arab Emirates and the United Kingdom
  • Latin America: Brazil, Colombia and Mexico
  • North America: Canada and the United States

DoubleVerify puts the margin of error for the marketer survey at 2.2 percentage points, at 95% confidence. Global averages are "weighted by region," according to the report, which says the year-on-year comparisons rely on "a similar survey in 2025." PPC Land's coverage of that earlier edition recorded 1,970 marketers interviewed in March 2025.

The methodology leaves a good deal unsaid. It gives no definition of a decision-maker by job title, company size or media budget, and no split between brands and agencies. Most charts carry no respondent base. The chart of concerns about AI optimization does not print the question that produced it, and none of the bidding charts states how many people answered.

The report also sets out its own vocabulary. Agentic AI, it says, "can autonomously make decisions and execute complex tasks, such as adjusting a bid to optimize toward a specific KPI." Third party, in DoubleVerify's usage, means AI providers that operate outside demand-side platforms, the systems brands and agencies use to buy programmatic inventory. The survey covers both the AI features built into DSPs and walled gardens such as Google and Meta, and independent providers such as DoubleVerify, according to the report.

That definition of agentic AI is wide. An automated bidding algorithm chasing a cost-per-acquisition target fits inside it, which helps explain how the report reaches such high usage figures.

Near-universal use, by DoubleVerify's measure

The adoption numbers are high. According to the report, 97% of marketers expected to use AI for media planning and buying in 2026, 92% use agentic AI for media planning and buying, and 77% describe themselves as "somewhat to very confident" in agentic AI.

The wording around those numbers is not consistent. The report states that 92% "use" agentic AI. The email that announced it says "nearly all use or plan to use agentic AI to some extent," a looser formulation that folds intentions into the count. Neither document prints the question behind the 92% figure, and the confidence figure combines two answer categories without showing how many respondents picked "very."

DoubleVerify's own summary is more restrained than its headline statistics. "Marketers haven't yet handed over all media planning, buying and optimizing to AI," the report states. "But this year, they're closer to doing so than ever before."

The task list

The report's most detailed chart tracks seven steps in the planning and buying process over three years. Respondents were asked whether they used AI at each step "last year" and whether they planned to use it "this year," so the 2024 column comes from the 2025 survey and the 2025 column from the 2026 survey. Creative work through dynamic creative optimization, or DCO, sits in the middle of the pack.

Task Used in 2024 Used in 2025 Planned for 2026
Analyze data 43% 55% 52%
Build media strategy 45% 44% 45%
Optimize creatives (DCO) 37% 39% 42%
Summarize media brief 28% 37% 40%
Activate campaign 31% 35% 40%
Optimize mid-flight 21% 31% 36%
Optimize bidding 24% 29% 34%

Data analysis is the most automated step and the only one where planned use for 2026, at 52%, sits below actual use in 2025, at 55%. "It appears that marketers still want or need humans to crunch the numbers," according to the report. Building a media strategy dipped from 45% to 44% between 2024 and 2025, the only step to fall in that period, before plans for 2026 returned it to 45%.

Every other step climbed in both years. Between 2024 and 2025, data analysis rose fastest, by 12 points, followed by mid-flight optimization at 10 points and brief summaries at 9. None of the seven steps, however, is handled by AI for a clear majority of marketers, and only one passed the halfway mark in 2025.

Plans versus outcomes

Set against the 2025 edition, the chart offers a rare check on how marketers' AI plans turn out. Last year's report, as PPC Land recorded it, showed campaign activation rising from 31% in 2024 to a planned 41% for 2025. The 2026 report confirms the 31% starting point but puts actual use in 2025 at 35%. On that comparison, which is PPC Land's and draws on two different samples, campaign activation fell six points short of what marketers had planned. The 2026 plan, at 40%, is close to last year's.

Other rows cannot be matched. The 2025 edition grouped bidding and mid-flight optimization together, at 41% in 2024 rising to a planned 46%. The 2026 edition splits them into 24% and 21% for the same year. DoubleVerify does not explain the change, so the two editions cannot be compared on either step.

Bidding crosses the halfway mark

A separate question deals with outside bidding tools. Half of marketers now use third-party AI bidding solutions, against 42% in 2025, according to the report. The share planning to adopt them fell from 49% to 41%, while the 10% with no plans did not move.

Third-party AI bidding 2025 2026
Yes 42% 50%
No, but plan to 49% 41%
No, and have no plans to 10% 10%

DoubleVerify reads this as momentum. "The proportion of marketers planning to use and actually using AI for bidding optimization traded places in 2026, indicating that the waiting room has begun to empty out," according to the report. The 2025 figures match those PPC Land reported in August 2025.

A different reading is also available. In the 2025 survey, 91% of marketers either used third-party AI bidding or intended to. A year on, 50% use it. That calculation is PPC Land's, across two samples, and does not track individual respondents, but it suggests the waiting room is emptying slowly.

Latin America leads, North America trails

Adoption varies by region. Latin America leads at 58%, followed by Asia Pacific at 53%, Europe, the Middle East and Africa at 46%, and North America at 45%, according to the report's regional breakdown. North America, home to the world's largest advertising market, therefore has the lowest share of marketers using outside bidding tools. The regional chart is labeled "Third-Party AI or Automated Bidding," a slightly broader description than the "third-party AI bidding solutions" in the global question.

Two charts, two answers

The bidding figures sit awkwardly beside the task list. In the task chart, 29% of marketers used AI to optimize bidding in 2025 and 34% plan to in 2026. In the bidding chart, 50% use third-party AI bidding solutions. Third-party tools are, by the report's own definition, a subset of AI bidding, so the subset is larger than the whole.

The same gap existed a year earlier: 42% used third-party AI bidding in the 2025 survey, while the task chart records 24% using AI for bidding optimization in 2024. The questions are worded differently, and respondents may well have understood "AI" in the task question to mean generative tools rather than bidding algorithms. The report offers no reconciliation, and readers looking for a single figure for AI in bidding will find two.

What marketers expect from AI bidding

Marketers who use AI bidding want everything at once, according to the report. Six KPIs fall within three points of each other:

  • Minimize cost: 44% in 2026, against 45% in 2025
  • Maximize conversions: 44%, against 39%
  • Maximize impressions: 44%, against 41%
  • Maximize conversion value: 42%, against 41%
  • Optimize ROAS: 42%, against 45%
  • Maximize clicks: 41%, against 42%

"Which of these KPIs is their biggest priority? All of them," the report states, adding that this is "a reasonable expectation for technology that's built for performance and never gets tired." The largest movement was a five-point rise in conversions as a goal; return on ad spend fell three points.

98% improvement, and the question behind it

The report's boldest claim is that 98% of marketers who use AI bidding reported performance improvements. The question that produced the breakdown was "How did AI bidding solutions improve campaign performance last year?" - wording that presumes an improvement took place. The report does not show whether "no improvement" was an available answer.

The breakdown itself is less emphatic. No single improvement was reported by more than 44% of respondents:

Improvement Saw in 2025 Expect in 2026
Time savings and operational gains 44% 43%
ROAS 37% 38%
Audience precision and personalization 36% 49%
Conversion rates 34% 40%
CPA 34% 36%
Campaign scalability 28% 27%

Audience precision shows the widest gap between experience and expectation. Only 36% saw it improve in 2025, yet 49% expect it to in 2026, which the report calls "the runaway lead in expected gains." Time savings went the other way: it was the most common improvement reported for 2025, but slightly fewer marketers expect it in 2026. "AI bidding did pretty well against these KPIs," according to the report. All of the figures are self-reported. The report contains no measured lift data for the survey population.

What buyers weigh, and what worries them

Accuracy matters most when marketers evaluate AI optimization tools. According to the report, 35% consider the accuracy and effectiveness of recommendations, 29% data privacy and security compliance, 29% cost or return on investment, 27% transparency of AI decision-making, 27% brand suitability or compliance features, 26% ease of use, 25% integration with existing platforms, 24% vendor reliability, 23% flexibility and 20% scalability across campaign sizes.

The concerns run on a different scale. Three issues tie at 60%: transparency into optimization decisions; data quality, availability or privacy; and brand suitability or compliance. Trust in outcomes follows at 56%, cost at 55%, and integration, lack of internal expertise and measuring ROI at 54% each. Vendor lock-in comes last, at 49%.

Two things stand out. The first is that brand suitability, DoubleVerify's core verification business, appears near the top of both lists. The second is the distance between confidence and concern. Some 77% of marketers say they are at least somewhat confident in agentic AI, while six in ten worry about how its decisions are made. The report does not cross-tabulate the two, so it is impossible to tell whether the same respondents hold both views.

The time problem

DoubleVerify is candid about the central disappointment. "A big concern: The AI era is supposed to usher in game-changing time savings in media planning and buying," the report states. "AI has saved time within individual workflow steps. But overall, the total number of hours marketers spend on manual optimizations hasn't budged."

The distribution confirms it. In both years, 57% of marketers spent between 8 and 16 hours a week on manual optimizations. The share at 4 to 8 hours edged up from 28% to 29%, the share above 16 hours slipped from 9% to 8%, and 6% spent four hours or less in both years.

Measured as a global average against the 2025 edition, which described the figure as time spent on routine optimization tasks, the change is small. Marketers spent 9 hours and 58 minutes a week in 2026, against 10 hours and 12 minutes in the 2025 survey. That is a reduction of 14 minutes, or about 2.3%, in PPC Land's calculation. Across a 40-hour week, the 2026 figure equals roughly a quarter of working time, close to the 26% that the 2025 report cited.

Region by region

The regional figures, compared with the 2025 edition as PPC Land reported it, show where the minutes went:

  • Asia Pacific: 10 hours 29 minutes, down from 11 hours 1 minute, a fall of 32 minutes
  • Latin America: 10 hours 48 minutes, down from 11 hours 15 minutes, a fall of 27 minutes
  • Europe, Middle East and Africa: 9 hours 35 minutes, down from 9 hours 38 minutes, a fall of 3 minutes
  • North America: 9 hours 16 minutes, up from 9 hours 14 minutes, a rise of 2 minutes

The two regions with the highest adoption of third-party bidding, Latin America and Asia Pacific, recorded the largest falls. North America, the lowest adopter, recorded none. The pattern is suggestive rather than conclusive: the samples differ between years, the regional sub-samples are smaller than the global one, and the report does not test the link.

Other surveys point in the same direction. Mediaocean's 2026 H2 Market Report, based on 312 marketing professionals surveyed in May 2026, found that only 19% believe AI is causing a major workflow transformation, down from 28%. MiQ's survey of 3,169 marketers in 16 countries, released in November 2025, found that 72% planned to use AI in more ways but only 45% felt confident applying it.

Walled gardens: 88% present, 32% optimized

Social and video platforms are where the manual workload is most visible. According to the report, 88% of marketers advertise on social platforms, the same share PPC Land recorded from DoubleVerify's 2025 survey. Of those, only 32% use algorithmic optimization solutions for their campaigns.

Each platform is "a separate entity, with its own rules, metrics, optimization levers and reporting," so marketers "must go into each one individually to make similar adjustments," according to the report. The top challenges when advertising on multiple platforms:

  • Different optimization levers on each platform: 48%
  • Budget reallocation complexity: 41%
  • Lack of unified cross-platform reach: 41%
  • Lack of unified reporting: 38%
  • Inconsistent metrics and naming conventions: 38%
  • Duplicative manual work: 37%

The report does not say whether the 32% includes the platforms' own automated buying tools or only outside software. That distinction matters. A Taboola-commissioned survey of 200 senior performance marketers in the US and UK, covered by PPC Land on May 14, 2026, found 91% using Google's Performance Max at scale, against 36% using open-web campaign tools at scale. If the 32% refers to third-party optimization, the figure says little about how much automation marketers already run inside each platform.

What DoubleVerify sells

The last section of the report is a product description. DoubleVerify groups its verification, optimization and measurement tools under the DV Media AdVantage Platform, or DV MAP, and describes three components relevant to the survey.

DV Scibids AI "builds custom algorithms around a brand's own KPIs and first-party data, such as offline sales and attention metrics, and deploys them at scale," according to the report. DoubleVerify positions it against the bidding built into buying platforms: "The standard bidding optimization built into programmatic platforms often pushes every campaign toward cookie-cutter outcomes." The report gives no evidence for that claim. When DoubleVerify launched its DV Neura AI engine on June 17, 2026, it said Scibids optimized 25 billion impressions a month.

DV Neura is described in the report as "the fully embedded cognitive engine that powers DV's platform," trained "on 18 years of proprietary data from the nearly 10 trillion digital advertising transactions we process yearly."

DV Authentic AdVantage, which DoubleVerify introduced in June 2025 for walled garden video platforms, combines pre-bid controls, dynamic optimization through Scibids and post-bid measurement. DV Rockerbox multi-touch attribution, along with viewability, brand suitability and attention metrics, feeds the optimization, according to the report.

Suntory

The one case study involves Suntory, which ran DV Authentic AdVantage in four social and video platform campaigns "seeking high-quality reach at scale." In all four, according to the report, "costs dropped, brand suitability remained high, and ad awareness improved significantly." The headline figure is an increase of up to 30% in an "Ad Awareness Score" against a control group.

"Up to" means the best of four results. The report does not name the platforms, the markets, the campaign dates, the spend, the size of the cost reduction or how the awareness score was measured. That follows a familiar pattern. When DoubleVerify reported earlier this month that Weight Watchers' ChatGPT Ads cost per acquisition ran 15.3% below benchmark, the release likewise gave no spend, dates or benchmark definition.

The ownership backdrop

The report arrives with DoubleVerify in the middle of a sale. Nielsen agreed on August 6, 2026 to acquire the company for $13.60 a share in cash, an enterprise value of about $2.15 billion, with closing expected by the first quarter of 2027, subject to a shareholder vote and regulatory approvals. The report does not mention the deal.

PPC Land's analysis the following day questioned what independence means when the dominant US television audience currency also owns a verification vendor that checks the same impressions. That article cited second-quarter 2026 revenue of $193.8 million, up 3% year on year. The AI report sharpens a related question: DoubleVerify now measures media quality, sells the algorithms that buy against those measurements, and publishes the survey that tracks how well such algorithms perform.

Why it matters for marketers

The finding with the widest implications is also the least flattering to the industry, DoubleVerify included. AI is in use at every step of the planning and buying chain, and half of marketers have added an outside bidding layer on top of the tools built into their platforms. Weekly hours spent on manual optimization are almost exactly where they were a year ago.

DoubleVerify's explanation is fragmentation: each platform has its own levers, reporting and naming conventions, so automation within one system does not reduce the work of coordinating many. The report presents unified, cross-platform optimization as the answer, and it happens to sell one. Other explanations are possible and go untested. Checking an algorithm's work may itself be manual. New tools may add tasks as quickly as they remove others. Marketers may also have reinvested saved hours in more campaigns rather than fewer hours.

The concerns data adds a second layer. Transparency, data quality and brand suitability each worry 60% of respondents. Those are the same issues PPC Land has tracked through the year, from IAS promising campaign insights from an agent without an AI black box to IAB Europe's October finding that 86% of organizations use AI for marketing while only 48% have advertising-specific rules. Evidence on AI buying in live auctions remains thin, though DataBeat's network data showed agentic buyers paying $6.13 CPMs against $6.95 for conventional programmatic buyers while entering 86% fewer auctions.

The numbers deserve care. The sample is large and the margin of error modest, but decision-makers are not defined, question wording is missing for several headline figures, and two charts disagree on how many marketers use AI for bidding. The 98% improvement figure rests on a question that assumes the improvement. And the report is published by a vendor whose product pages sit alongside the data. None of that makes the 9 hours and 58 minutes less telling. A company that sells automation has reported that automation has not yet given marketers their time back.

Timeline

Summary

Who: DoubleVerify, the ad verification and optimization company that Nielsen agreed to acquire in August 2026, with fieldwork by Sapio Research among 2,020 marketing and advertising decision-makers in 25 countries.

What: A report titled "Beyond the Hype: The Real State of AI in Digital Advertising" finding that 97% of marketers expected to use AI for media planning and buying in 2026, 50% use third-party AI bidding (up from 42%), and marketers still spend 9 hours and 58 minutes a week on manual optimizations, against 10 hours and 12 minutes in the previous survey. The report also contains promotional material for DoubleVerify's DV Scibids AI, DV Neura and DV Authentic AdVantage products.

When: The survey was conducted in February 2026. DoubleVerify distributed the report by email on October 29, 2026.

Where: The sample spans Asia Pacific, Europe, the Middle East and Africa, Latin America and North America. Latin America shows the highest use of third-party bidding at 58% and North America the lowest at 45%.

Why: The figures test the industry's central promise for AI in media buying, that it saves time. By DoubleVerify's own account, the total hours have not budged, while concerns about transparency, data quality and brand suitability each affect 60% of marketers. The report's methodological gaps, including undefined respondents, missing question wording and conflicting bidding figures, and its publisher's commercial interest in AI optimization, limit how far its conclusions can be taken.