Advertisers Love AI Until It Reaches For The Budget
Artificial intelligence has becomenearly universal in advertising, but marketers still aren't ready to let the machines wander unsupervised through their campaigns with a corporate credit card.
A new StackAdapt report found that 91% of marketers use AI tools in marketing or advertising. About 86% usethem regularly or for most tasks. Another 88% reported some form of AI-driven performance improvement.
The enthusiasm cools when AI moves from summarizing reports to making decisions.Advertisers are happy to let it analyze performance, draft creative and perform the chores that make humans question their career choices. They're much less comfortable letting it controlbudgets, adjust bids or make consequential campaign changes without supervision.
StackAdapt calls this divide the "AI delegation gap." The report is based on a NewtonX survey of500 marketing and advertising professionals at mid-sized and enterprise organizations. It also includes a separate survey of 187 StackAdapt clients.
The Robot May Recommend, But The HumanStill Approves
Advertisers aren't rejecting automation. They're drawing boundaries around how much authority it gets.
About 90% of respondents agreed that AI shouldrecommend actions while humans make the decisions. Roughly 89% supported allowing AI to prepare actions for human approval. Support fell to 78% when AI would act within rules set by humans.
Only 50% agreed that AI should operate autonomously after proving its performance.
In other words, advertisers want a highly capable assistant, not a junior employee who discovers the"increase budget" button at 2 a.m.
The caution becomes more pronounced among the people who actually manage campaigns. About 58% of strategic decision-makers were comfortable withautonomous AI, compared with 47% of hybrid operators and 34% of hands-on practitioners.
That gap won't shock anyone who's watched leadership announce a bold technology initiativeand then hand the cleanup to the people closest to the work. Executives see speed, scale and efficiency. Practitioners see pacing errors, brand-safety violations and a client asking why an algorithmspent $40,000 before lunch.
Generic Recommendations Are Getting The Treatment They Deserve
Only 6% of marketers said they almost always act on AI recommendations deliveredthrough advertising platforms. The biggest reason for ignoring them was painfully simple: 42% said the recommendations felt generic or irrelevant to their campaign.
Another 22% said theguidance didn't align with their strategy. About 17% cited a lack of explanation or transparency. Meanwhile, 12% suspected the recommendation was timed to increase spending rather than improveperformance.
Imagine that. Advertisers have noticed that a platform earning money from ad spending occasionally recommends more ad spending. Somewhere, a "boost budget"notification is pretending to be deeply offended.
Recommendations received positive ratings of no more than 40% across relevance, timing, clarity, manageability and ease of action. Explanationand reasoning ranked lowest at 31%.
Marketers were most likely to act when a recommendation included a clear rationale, cited by 33%, or connected directly to a meaningful key performanceindicator, cited by 31%.
The lesson for platforms is fairly brutal. Slapping "AI-powered" on a generic suggestion doesn't transform it into strategic counsel. It just givesstale advice a more expensive costume.
Reporting Is AI's Safe Space
AI has gained the most traction in areas where it can help without being allowed to detonate the mediaplan.
About 77% of respondents use AI for reporting and summaries while 74% use it for performance analysis and insights. Creative development followed at 69% and research and audiencediscovery came in at 68%.
Usage fell to 51% for optimization and in-flight adjustments. Campaign setup and budgeting and bidding each stood at 40%.
Advertisers also want AI to becomemore proactive. About 41% want it to flag risks before they escalate and the same percentage want it to summarize insights. Another 39% want recommendations tied to campaign goals while 35% want AI toexplain why performance changed.
That's a useful progression. First, AI tells buyers what happened. Then it explains why. Eventually, perhaps, it'll reveal why the campaign has 17naming conventions and three people claiming ownership of the same conversion event.
Still, AI-generated insights can sound certain even when they're missing crucial information. About34% said recommendations don't reflect campaign goals or context while 37% cited AI's limited ability to explain its reasoning.
Confidence remains a feature. Fluency isn'tproof.
Brand Risk Remains The Giant Red Button
Brand risk was the biggest obstacle to delegating more authority to AI, cited by 63% of respondents. Data quality concernsfollowed at 56%. Lack of transparency and performance volatility each came in at 36% while 34% cited poor integration between systems.
Advertisers want several safeguards before loosening theleash. About 37% want to see a confidence level or predicted impact before applying a recommendation. Another 36% want limited-budget testing while 35% want a clear explanation of why a change wasmade.
Full audit trails and easy reversal options each were cited by 31%.
These aren't unreasonable demands. If AI can change bids, budgets or targeting, marketers would like toknow what it changed, why it changed it and how to change it back before the postmortem becomes a recurring calendar invitation.
Defaults also may be quietly expanding AI's authority.Only 22% said they intentionally selected most of their platform's automation features. About 38% said most choices were deliberate but some defaults were accepted without review. Another 29%described the mix as roughly half deliberate and half default.
The remaining 11% said most active automation features were platform defaults or unclear by default. That's less"strategic AI adoption" and more "someone clicked continue during setup."
Efficiency Is Winning, But Revenue Still Wants Receipts
AI's most visiblebenefits are operational.
About 62% of respondents reported spending less time on manual optimization. Some 56% cited faster campaign setup or launch while 53% reported faster optimizationcycles. Better audience targeting was cited by 47%.
The numbers were less impressive for direct business results. Only 27% reported improved return on ad spend and 22% cited lower cost perclick.
That doesn't mean AI isn't creating value. It means marketers have stronger evidence that it saves time than that it makes money. The industry has successfully demonstratedthat AI can help people finish tasks faster. Finance departments may eventually ask whether any of those tasks improved the business.
StackAdapt also cautioned that its findings can'tprove deeper AI usage causes better results. Higher-performing organizations may simply be more willing to adopt AI broadly.
Leadership's AI Ambition Meets The Plumbing
About 79% of respondents feel moderate or strong pressure to increase AI usage. Yet 59% said leadership's expectations may be running ahead of the organization's readiness.
Only19% said their AI tools are fully integrated into marketing and advertising workflows. Nearly half cited fragmented data pipelines while 41% said customer relationship management or first-party dataisn't connected with buying platforms.
Meanwhile, 83% said inconsistent AI capabilities disrupt efficiency at least somewhat.
The pressure is coming largely from executives.About 35% called C-suite leadership an extreme source of pressure to increase AI usage. Competitive anxiety and industry trends followed at 28%.
Apparently, announcing an AI mandate is easierthan connecting the data, defining approval rules and deciding who gets blamed when the machine produces an expensive surprise.
Accountability is already murky. About 31% said the teamcollectively would be responsible for poor AI-driven decisions. The same percentage pointed to leadership or the person who approved the recommendation. Another 28% named the campaign manager while26% chose the individual operator or campaign owner. Respondents could select multiple answers.
Most concerning, 17% said there's no clear accountability at all.
AI may be movingquickly, but responsibility still appears to be buffering.
Trust Will Determine How Much AI Gets To Do
StackAdapt's findings suggest advertisers will delegate moreauthority when AI is relevant, transparent and controllable. They also want recommendations that can be tested, reversed and traced to an accountable owner.
That puts the burden on platformsto prove that their AI understands the campaign instead of merely having access to it.
Advertisers aren't demanding perfection. They're asking for context, guardrails and evidence.Until platforms deliver those things, AI will remain welcome at the marketing table. It just won't be handed the company wallet.
