DoubleVerify: third-party AI bidding reaches 50%, but manual work barely shrinks.
Marketers still spend almost ten hours a week adjusting campaigns by hand, according to a new DoubleVerify survey, even though half now use outside AI tools to set their bids. That is only 14 minutes less than a year ago.
The report, "Beyond the Hype: The Real State of AI in Digital Advertising", went out to marketers by email on 29 October, PPC Land reported on 1 November. Sapio Research ran the fieldwork in February 2026, surveying 2,020 marketing and advertising decision-makers in 25 countries, Saudi Arabia and the UAE among them, with a stated margin of error of 2.2 points. Average time on manual optimisation came to 9 hours 58 minutes a week, against 10 hours 12 minutes in the previous survey. Use of third-party AI bidding rose from 42% to 50%, while the share planning to adopt it fell from 49% to 41%.
Adoption claims are high elsewhere too. The report says 92% use agentic AI in planning and buying and 77% are at least somewhat confident in it, yet 60% worry about how optimisation decisions are made, the same share as for data quality and brand suitability. In Europe, the Middle East and Africa, 46% use third-party bidding, and manual time fell by just three minutes, to 9 hours 35 minutes. Some 88% advertise on social platforms, but only 32% use algorithmic optimisation there.
There are reasons for caution. DoubleVerify sells its own AI bidding product, DV Scibids AI, so it has a stake in the result. The report does not define a decision-maker and prints no respondent base on most charts. PPC Land also points out that another chart in the same report puts AI use for bidding at 29% in 2025, below the 50% for third-party tools alone, with no explanation.
For Gulf teams buying across Meta, TikTok, Snap and Google at once, the useful finding is the cross-platform one: the challenge cited most often, by 48%, is that each platform has different optimisation levers. Automation inside each platform hasn't removed that work, so ask vendors for measured time savings rather than survey claims.
Why it matters
AI bidding is spreading fast, but the time saving it promises hasn't arrived yet, and the figures come from a company that sells the tools.
Related trends.
S&P moves Meta's outlook to negative over AI data centre spending
S&P Global Ratings has revised its outlook on Meta Platforms to negative from stable while keeping the company's AA- credit rating. The agency warns that leases for AI data centres could push Meta's debt measure past its downgrade threshold as early as 2027.
Why it mattersMeta's AI bill now carries a credit warning, which raises the pressure to turn that spending into ad revenue through more automation and new ad products.
Funnel survey: 36% of marketers keep funding channels they doubt
More than a third of in-house marketers say they have kept paying for a channel they believed wasn't working because its numbers looked too good to cut. The finding comes from Playing defense, published on 29 October by Funnel and Ravn Research.
Why it mattersIt puts numbers on a familiar problem: when teams can't prove what works, budgets drift to whatever is easiest to defend, not what performs best.
du wins two regional awards for its brand evolution and architecture
du won Brand Evolution at The Marketing Society GCC Awards 2026 with TBWA\RAAD, and Gold for Best Brand Architecture Solution at the Transform Awards MEA 2026 with Landor. Communicate reported both wins on 30 October.
Why it mattersTwo awards recognise one telecom rebrand, on positioning and on structure. UAE case; results come from a single report.
