A consumer brand should use AI first to measure what its marketing returns, on measures finance signs before each campaign launches, and hold AI-made ads and emails until that measurement is running. Measurement happens where customers never see it, and once it runs, a test of AI-made ads can be read in sales like any other campaign.
Gartner's release of March 16, 2026 reports that half of US consumers say they would prefer to give their business to brands that do not use GenAI, where a brand that uses GenAI is one that puts it in consumer-facing messages, advertising and content. The release gives no sample size or fieldwork date for that figure. It also reports a survey of 1,539 US consumers, conducted in October 2025, in which 68 percent frequently wonder whether the content and information they see is real. In the same release, a Gartner analyst advises marketers to start with clearly assistive use cases that deliver immediate customer value.
In a survey by Bain and Google of almost 1,400 senior marketing and finance executives worldwide, published in July 2026, only 41 percent of marketers feel appropriately equipped with the right data, tools, and measurement capabilities to tie their performance to business outcomes, and some 19 percent of marketers surveyed have mature AI capabilities for analytics and insights. Bain writes that the leaders in its research lock metrics before campaigns begin.
For a maker of packaged snacks selling through grocers and its own website, the CFO and the head of marketing sign three entries before each campaign goes live: the number the campaign is meant to move, such as retail sales from point-of-sale data in the regions where it runs; what that number will be read against, for example matched regions without the campaign; and the date the result gets read. AI can do the analysis behind the read: choosing matched regions from past sales, pulling media spend, retail sales and web orders into one set of figures each week, and producing the comparison on the agreed date.
The measurement is running once the first campaign has been read that way on its date, so that date is also the earliest a test of AI-made ads or emails can start. The test is read the same way, against regions or customers that got the brand's usual content. Ads and emails whose words or images AI generated count as AI-made, while resizing or translating the brand's own work can carry on.
The three signed entries also give the board of a PE-backed brand one set of marketing numbers that finance has agreed. In the Bain and Google survey, companies with strong marketing and finance relationships are 2.5 times more likely to have more credible data, comparing 501 respondents with strong relationships against 138 with weak ones.
