AI has made marketing content cheap to produce. It has not made it more effective. Most marketing leaders can now feel the gap: the team ships more assets every quarter, and each one does less.
The numbers say the feeling is widespread. Dentsu Creative's 2026 CMO report, a survey of 1,950 senior marketing decision-makers across 14 markets, found that 76% worry they are producing more content for less impact or visibility, and 74% agree that more assets do not mean more impact. At the same time, 81% expect to need far more content going forward, and 71% say they do not know how to balance the cost of that content against the return it generates.
The savings have not shown up either. In the same survey, 70% of CMOs said they have not yet seen major cost efficiencies from AI, even though 71% said it has made them faster to market. 74% are considering replacing some of the tools they pay for today, and 62% said AI-assisted creative cannot yet capture their brand's tone of voice, look, and feel.
The volume problem also reaches the media plan. Research from WARC and CreativeX cited in the report found creative output up an average of 30% while media spend per asset fell 15%, and more than 90% of assets received less than $10,000 in media support. More pieces are competing for the same budget, so each one is seen by fewer people.
Volume was never the goal
When an asset was expensive to make, the cost forced a decision about whether it deserved to exist. AI removed that check. Now the default is to make the extra version, the extra post, the extra landing page, because making it costs almost nothing. The real cost moved to places the production budget does not show:
- Media spread across so many assets that none reaches enough people to work.
- Reviewers checking more drafts and catching less.
- A brand that starts to look and sound like every other company using the same tools. In the Dentsu survey, 80% of CMOs said that without strong craft up front, AI-enabled production risks more sameness, in more places, more quickly.
Judge content by what it moves
- Count outcomes, not output. Replace assets shipped and cost per asset with pipeline influenced, conversion on the pages the content supports, and how often sales actually uses it.
- Decide what deserves to exist before you make it. Every new asset needs a named audience, a job in the buying journey, and a plan to get it seen.
- Fund fewer pieces properly. Put media behind the assets that earn it instead of spreading thin across everything that ships.
- Retire what does not work. Each quarter, cut or merge content that no one sees or uses, the same way you would cut an underperforming campaign.
- Spend the time AI saves on craft. 82% of CMOs in the survey said production savings let them reinvest in upfront craft and creativity. Put that time into the idea, the proof, and the customer voice that make a piece worth reading.
Three questions for the next content review
- Which ten assets from last quarter moved pipeline, and what did they have in common?
- How much media support did the average new asset get, and was it enough to be seen?
- If we made half as much next quarter, what would we stop, and would any buyer notice?
This week
Pull last quarter's content list and sort each piece by the outcome it was meant to drive. Mark the ones that drove it, the ones nobody saw, and the ones no one can measure. The last two groups are your 2027 cut list, and the budget they free up belongs to the first.
