TripleLift this month published a global research report on how advertising professionals are deploying AI across campaign workflows, finding a sharp split between tasks where the technology has won over practitioners and those where human involvement remains essentially non-negotiable. The report, titled "The Evolution of AI in Global Advertising," draws on a survey of 200 advertising industry respondents conducted in April 2026 across six countries. It was released on May 19, 2026.

The findings capture an industry caught between appetite and apprehension. Most companies now run AI tools in some part of their campaign operations, yet fully autonomous execution remains the exception rather than the norm - and the closer a task gets to the creative act, the more practitioners pull back.

Wide adoption, shallow roots

According to TripleLift, 62% of respondents report that their company is still in the exploration or pilot phase of AI integration. That figure sits alongside a separate finding that 60% of companies say they have a centralised AI strategy. The gap between having a strategy and having confidence in it is significant: fewer than 30% of respondents express high confidence in that strategy. Only 6.5% describe themselves as "extremely confident," while 49.5% fall in the "somewhat confident" band.

These numbers suggest that AI has moved from curiosity to operational fixture at many organisations, but that the transition is shallow. Teams are running pilots and early deployments, not yet institutionalised workflows with clear governance and performance benchmarks. The sample of 200 respondents answered up to 30 questions each, spanning buyers, planners, and technologists across the United States (113 respondents), the United Kingdom (34), France (31), Canada (17), Germany (3), and Australia (2).

The four execution pillars - and the gap

TripleLift frames campaign execution around four distinct pillars: media, measurement, audience, and creative. Each requires a different type of AI system. According to the report, the three most prevalent AI pathway combinations in use today link media and measurement together, or media, measurement, and audience in tandem, or measurement and audience without media. In all three cases, the creative pillar is absent.

The report describes creative as "the absent link," noting that it demands the highest level of human intervention of any of the four pillars. Where the other three can be connected through integrated AI systems to allow near-continuous campaign setup - requiring human approval rather than end-to-end manual crafting of deals or assets - creative remains a bottleneck that AI has not crossed.

The data tells a specific story

The adoption numbers across those four pillars are striking in their unevenness. According to TripleLift, 73% of respondents currently use AI for campaign optimization, covering bidding, audience optimisation, and creative asset refinement. That is the highest adoption rate of any task category in the survey. Audience targeting comes second at 59%, with AI being used for audience segmentation and the construction of contextual and behavioural targeting segments.

Creative production, by contrast, sits at 25%. Even within that 25%, the pattern is not wholesale automation: over half of those using AI in the creative space deploy it to scale creative tests and adjust existing assets, while fewer than half are generating entirely AI-produced outputs. The implication is that AI in creative is primarily an assistant to human production rather than a replacement for it.

The gap between optimisation at 73% and creative production at 25% is one of the most concrete data points in the report. It maps directly onto where advertisers draw the line between trusting a machine and insisting on human judgment.

Full automation of campaign execution is rarer still. Only 19% of respondents report that AI fully automates their campaign execution process. At the other end, 40% still maintain manual control across the entire execution workflow. Somewhere between those poles, 38% say AI fully automates audience targeting specifically, while 31% indicate that budget allocation is handled entirely by AI. Creative and supply-side levers, in contrast, still require direct human involvement in most organisations.

The review tax

One of the more concrete and quantifiable findings in the TripleLift report concerns what the company calls the "review tax" - the time advertising professionals spend checking AI-generated outputs before they go live. The time saved by AI on execution does not simply return to practitioners as free hours. According to the report, 45% of respondents spend between one and four hours per week reviewing AI outputs. A further 22% spend between four and eight hours on the same activity. At the high end of the scale, 6% dedicate more than 12 hours per week to ensuring that campaign materials meet brand standards.

That final figure is worth pausing on. Twelve hours a week reviewing AI outputs is, in practical terms, a significant share of a working week being allocated specifically to catching and correcting what the machine produces. Even for the majority group spending up to four hours weekly, the review process represents a non-trivial commitment - and one that compounds across teams and campaigns.

This dynamic is not unique to TripleLift's findings. PPC Land has documented growing advertiser unease with AI-generated content in programmatic environments, with 82% of programmatic buyers now citing AI-powered optimisation as essential when selecting partners but simultaneously flagging quality control as a persistent concern. The tension between capability and oversight is a recurring theme across the industry in early 2026.

Why trust is the limiting factor

When TripleLift asked respondents who maintain human intervention in their AI-driven workflows to explain why, the answers cluster around a specific set of concerns. Among the 190 respondents who answered that question (question 28 in the survey), 67% cited lack of trust in AI output as the primary reason. Technical errors and hallucinations came second at 61%. The need for brand safety checks registered at 41%, and data privacy concerns at 21%.

The 67% figure on trust in output is notable because it precedes the other concerns in frequency. Brand safety and technical errors are specific, addressable problems - the kind an improved model or better guardrails can reduce. A general lack of trust in the output is harder to resolve. It reflects a more fundamental scepticism about the reliability and judgment of AI systems in contexts where a mistake - the wrong image, an off-brand phrase, a contextually inappropriate placement - can cause reputational damage before anyone notices.

A January 2026 analysis on PPC Land identified precisely this pattern, documenting that advertisers face genuine operational uncertainty around AI creative quality rather than simply technical friction. The TripleLift data provides a numerical anchor for that observation.

Where AI is already delivering

The report does not present a uniformly cautious picture. In the areas where AI has found footing - primarily data-driven functions rather than expressive ones - practitioners are seeing measurable improvements. According to TripleLift, early adopters report gains across three categories: speed (covering both campaign launch time and optimisation cycle duration), efficiency (measured through eCPM and cost-per-click), and media metrics including click-through rate and conversion rate.

These are the areas where AI's numerical and pattern-matching strengths map most directly onto what advertising operations require. Bidding, for example, involves processing large volumes of auction data faster than any human team can manage and adjusting in real time. Audience segmentation requires correlating first-party and contextual signals across large datasets to identify likely responsive groups. Both are tasks where algorithmic systems have demonstrable advantages.

What remains more contested is whether those efficiency gains in the backend actually translate into better outcomes for brands whose value depends on how they communicate creatively. The TripleLift data suggests the industry's working answer, for now, is that they do not - at least not enough to relinquish human creative control.

The programmatic industry context

The timing of the TripleLift report lands in a period of concentrated activity around agentic AI across the programmatic supply chain. TripleLift itself announced its TL Spark agentic intelligence layer in April 2026, positioning it as an orchestration system linking creative, supply, audience, and measurement rather than automating any single function independently. The architecture described in that announcement reflects the same four-pillar framework the research report maps out.

Elsewhere in the supply chain, PubMatic launched AgenticOS in January 2026 as infrastructure for autonomous campaign execution. The IAB forecasted 9.5% US ad spend growth for 2026, with agentic AI systems cited as a central driver. Mediaocean's 2026 H1 Outlook Report, released in January 2026, found that marketers plan to increase AI media spending at a higher rate than traditional search advertising - a first in the survey series - while simultaneously reporting difficulty implementing generative AI tools internally.

That last finding mirrors what TripleLift's survey documents at the campaign execution level. Organisations are allocating budgets toward AI-powered media channels while struggling to systematise AI use within their own workflows. The external AI environment is moving faster than the internal one.

The Criteo and TripleLift partnership, reported by PPC Land in April 2026, illustrates how platform-side AI integration is advancing regardless of advertiser-side hesitancy. That collaboration combines Criteo's Commerce Audiences with TripleLift's curation layer into activatable Deal IDs across web, mobile, and CTV - automating a portion of the supply chain that was previously manual. The advertiser may still be reviewing outputs manually, but the infrastructure beneath them is increasingly automated.

Survey methodology

TripleLift conducted the survey in April 2026. The 200 respondents answered a maximum of 30 questions each. Geographic distribution was weighted toward the United States (113 respondents, or 56.5% of the total), with the United Kingdom second at 34, France at 31, Canada at 17, Germany at 3, and Australia at 2. The report was produced with contributions from Alana Schept (Associate Director, Global Marketing), Andrea Strum (Research Director), and Emma Sainsbury (Research Analyst), among others.

The sample size of 200 is relatively small for a global industry survey, and the geographic skew toward US respondents is worth noting when reading any cross-country inferences. The results are indicative of attitudes within TripleLift's network and professional ecosystem rather than a statistically representative cross-section of the entire advertising industry.

Future priorities

According to TripleLift, the industry's wishlist for future AI use centres on four areas in ranked order: inventory discovery, deal and package curation, creative testing at scale, and AI adaptations to existing creatives. The ordering is telling. The highest-priority items are supply-side and transactional - functions closest to where automation already works. Creative remains last, even in the aspirational view. "AI adaptations to existing creatives" is a more limited ambition than fully AI-generated production, suggesting the industry envisions a future where AI adjusts and scales human-made creative rather than originating it.

Timeline

Summary

Who: TripleLift, a supply-side platform and Creative SSP, surveyed 200 advertising professionals across the United States, United Kingdom, France, Canada, Germany, and Australia as part of its 2026 AI research report.

What: The report "The Evolution of AI in Global Advertising" documents a wide adoption gap between AI use in campaign optimisation (73% of respondents) and AI use in creative production (25%), with only 19% of respondents fully automating campaign execution. It identifies a "review tax" where 45% of practitioners spend up to four hours per week reviewing AI-generated outputs, and finds that 67% cite lack of trust in AI output as the primary reason for maintaining manual oversight.

When: The survey was conducted in April 2026. The report was published on May 19, 2026.

Where: The survey covers advertising professionals across six countries: the United States (113 respondents), United Kingdom (34), France (31), Canada (17), Germany (3), and Australia (2). The findings apply across programmatic advertising contexts including open web, mobile, CTV, and retail media.

Why: The report matters for the marketing community because it quantifies the gap between where AI capability has reached and where practitioner trust lags behind. The creative pillar - the one most connected to brand identity and consumer perception - remains the least automated component of campaign execution, even as data-driven functions have reached high adoption. For practitioners navigating AI tool adoption decisions, budget allocation between human and automated production, and internal governance around AI workflows, the numbers provide a concrete reference point grounded in industry peer behaviour.