The gap between what brands spend on AI-influenced media and what they spend to measure it is now the most expensive unaccounted line item in marketing. IAB's September 2026 outlook confirmed U.S. ad spend growth hit 12.3% this year, up from a 9.5% forecast in January. Commerce media alone is projected to grow 13.6%. The measurement systems tracking that spend have not kept pace. Forty-four percent of the 211 U.S. brand and agency decision-makers surveyed cite adapting to AI-driven consumer behavior as their top media investment challenge, while 45% say comparing AI-driven and traditional customer journeys is a major measurement problem.
The companies pouring more budget than ever into AI-influenced channels are the same ones that cannot tell whether those channels worked. The forecast changed. The attribution stack did not.
The dashboard that lies to your CMO
Your multi-touch attribution dashboard still shows the last paid click. A buyer who discovers your brand through a ChatGPT answer, searches your branded term three days later, and converts through a direct visit shows up as a direct conversion. The AI influence vanishes. The report your CMO receives under-counts AI influence and over-credits display and search. The next budget cycle quietly shifts spend away from AI-influenced channels because the dashboard says they underperform. The channels are not underperforming. They are unmeasured.
Seventy-six percent of marketers now say their primary focus is optimizing content for AI-generated answers, and 72% are optimizing for AI LLM models. The money is moving into discovery layers that last-click attribution was never designed to read. Your stack was built for a web where the journey started with a paid click. That web is shrinking.
What marketers are patching together
Eighty-six percent of IAB respondents are changing how they measure media performance because of AI and agents, or expect to within twelve months. The stopgap solutions are practical but fragmented. Forty-eight percent are measuring brand visibility and citations in AI tools. Forty-four percent use branded search and direct traffic as proxies for AI-driven discovery. Forty percent are buying third-party AI discovery analysis tools. Thirty percent are running incrementality tests. Another 30% are turning to modeled measurement.
None of these solve the problem outright. All are better than the MTA dashboard that reports nothing about AI influence. The most telling number is what marketers are not doing: only 26% are putting less weight on website traffic. The old measurement stack is not being replaced. It is being layered over with new metrics nobody has integrated into a coherent system. Every team uses a different signal, every dashboard tells a different story, and every budget review ends in an argument about which proxy is least wrong.
The agent traffic you cannot classify
Twenty-seven percent of respondents list bots and agents outnumbering humans in web traffic as a media investment concern. Twenty-eight percent cannot distinguish humans from legitimate agents. Thirty-three percent cannot tell a legitimate agent from a bot or fraud. When an agent clicks your ad, browses your product page, and triggers a conversion event, your analytics record a session. You have no way to know whether that session came from a human evaluating your product, an agent acting on the human's behalf, or a bot generating fraudulent engagement your media budget paid for.
Teams that have addressed this are running separate measurement systems for human and agent traffic - agent-aware analytics that tag sessions by whether a real person or a declared agent initiated them. Teams that have not are optimizing spend against metrics that include a growing percentage of agent and bot traffic. The optimization is making their spend worse. They do not know it yet.
What to do in the next 90 days
First, audit the percentage of your web and conversion traffic that is non-human. If you cannot answer that question, you lack the data to optimize any AI-influenced media channel. The IAB findings suggest the number is large enough to invalidate last-click attribution for channels AI discovery touches.
Second, separate AI-influenced conversion paths from last-click conversions in your reporting. The buyer whose journey started with a ChatGPT recommendation and ended with a direct visit is not the same buyer who clicked your paid search ad first. Do not average them together. Report them separately, measure them separately, and budget against the measurement that matches the channel.
Third, pilot one incrementality test on an AI-influenced channel this quarter. Thirty percent of marketers are already doing this. The teams running incrementality tests against AI discovery will know by year-end whether their AI media spend is working. The teams that are not will still be arguing about it during the 2027 budget review. For teams looking to build the analytical skills to run these tests, AI Data Analysis Courses focused on measurement design and attribution modeling can close the capability gap quickly.
Why this matters for marketing and sales teams
Your media budget grew 12.3% this year. Your measurement budget did not. The gap between the two is now the largest source of unmeasured return in your marketing organization. For sales and marketing leaders, this means the performance data you present to finance is increasingly incomplete. The teams that close the measurement gap in the next two quarters will have the cleanest read on what works in an AI-influenced media environment. The teams that do not will spend 2027 explaining to their CFO why their AI media investment does not appear in the attribution dashboard - and why the budget requests keep rising for channels the numbers refuse to validate.
Your membership also unlocks:
