AI search is now the top media challenge, and buyers are answering it with brand spend

Adapting to AI-driven search is now the single biggest media investment challenge in US advertising, named by 44% of buyers in the IAB's 2026 Outlook Study September update, published on 10 September. The trade body now expects 12.3% growth in US ad spend this year, against the 9.5% it forecast in January. What the study does not show is money moving into answer-engine tactics. It shows money moving into brand.

The behavioural half of that story was reported five days later by Digiday, which found chief marketers using anxiety about AI search to argue internally for bigger budgets and more upper-funnel work rather than for optimisation projects. Alyson Griffin, who heads marketing at State Farm, said the shift in search behaviour raised rather than lowered the value of the insurer's sports-adjacent brand campaigns. "We value it. We're not pulling it back," she told Digiday, adding that the concern had become a live factor in the company's media budget debates.

Acquisition jumped. Repeat purchase did not move.

The priority numbers are where this gets interesting for anyone selling software. Customer acquisition is now a stated media investment goal for 63% of buyers, nine points higher than in January. Brand equity sits at 43%, six points up. Repeat purchases barely moved, holding at 24% — a priority the IAB says had almost doubled since 2024.

Read as a reflex, that is coherent: if a machine now stands between a category question and your name, the defensible position is to be the name people already carry into the question. IAB's Chris Bruderle framed the wider change as a move away from internal tooling questions. "Now the question is becoming, 'How do I reach a consumer who uses AI?'" he said in the release.

But the flat line on repeat purchase is the number a software marketer should stare at, because it inverts on contact with a subscription business. The brands driving that survey sell insurance policies, cars and running shoes — purchases separated by years, where acquisition genuinely is the revenue event. In subscription software the revenue event is the renewal and the seat expansion. Copying that budget shift means funding the one priority that matters least in the model software actually runs on.

It is also worth being precise about what caused the forecast upgrade, because the two sources do not quite agree. Digiday reports the IAB as crediting AI-search-driven spend increases as part of the lift. The release itself attributes the stronger first half to cyclical events — the Winter Olympics and the FIFA World Cup — plus easing worry about the macroeconomy, and treats the AI findings as a separate section on priorities and challenges. The AI-search story is a real shift in stated intent. It is not, on the evidence the IAB published, the thing that moved the number. A trade body for the digital advertising industry also has an obvious interest in a healthy forecast, and the study rests on stated priorities from just over 200 brand and agency decision-makers.

The measurement gap is the part that transfers

Eighty-six percent of buyers say conversational AI tools and agents have already changed their performance measurement, or will within the next twelve months. The most common response, chosen by 48%, is to measure brand visibility and citations inside the AI tools themselves. Forty-five percent say their hardest measurement problem is comparing AI-driven journeys with traditional ones.

"Your numbers do the talking… that can unlock investment from the people that hold the purse strings"

That line, from Mediaplus head of digital performance Oliver Williams speaking to Digiday, explains why citation tracking is spreading so fast: it produces a chart in a budget meeting. The problem is that a citation count measures the output of familiarity, not the cause of it, and it arrives without a baseline anyone trusts. Further down the IAB's list of responses sits a duller instrument: treating branded search volume and direct traffic as the proxy. For a software company that is the cheaper and better one, because you already own both series, they go back years, and nobody has to buy a tool to believe them.

What a software marketer should do differently

The transferable part is not the channel. Digiday's examples all resolve to mass-reach vehicles — linear television, connected TV, sports sponsorship — that a category with a few thousand real buyers cannot use. What transfers is the mechanism: fund whatever causes people to name you unprompted, then measure whether they do.

• Pull three years of branded-search impressions and direct-session counts before you commission any AI visibility tooling. That is your baseline, and it predates the thing you are trying to measure.
• Split the budget request by revenue event, not by funnel stage. If renewal and expansion carry your revenue, the acquisition-heavy reflex in the IAB data is the wrong template to argue from.
• Use the 44% finding as internal evidence, with the caveat attached. Presenting it as a measured revenue effect rather than a stated concern is the fastest way to lose the argument the second time you make it.