adtestbench Higgsfield
EN

AI news · Agencies · AI briefs

Agency chiefs say AI is bloating the briefs clients send

M&C Saatchi’s North American chief puts AI-written client briefs at 85%, and agency staff describe long documents with no clear goal.

By Alexandre S. , 08:40 UTC

Agencies say the briefs clients write with AI have grown longer and emptier. Nadja Bellan-White, who runs M&C Saatchi in North America, estimated at Digiday’s Media Buying Summit that 85% of the briefs her agency gets from clients were written using AI tools, Digiday reported on Thursday 5 November 2026. The result, Digiday writes, is documents hundreds of pages long, full of marketing jargon with little vision behind it.

Retro-futurist illustration: a striped sunset over a grid horizon under a starry sky, with a ringed globe standing on the horizon.
Drawn by adtestbench from “A little bit ridiculous: Agencies find AI overuse is leading to brief bloat”,

Other executives at the event, speaking under Chatham House rules, described briefs that never state the marketing goal. Bellan-White called the trend “a little bit ridiculous” and urged people to switch off their laptops and phones and talk face to face.

Among organisations that deploy AI agents, 56.4% use them to create campaign briefs, by a Basis Technologies survey of 184 professionals taken in August 2026. Basis’s release says 98.4% use generative or agentic AI at work. Claude is used by 72.3% of the marketers surveyed, up from 17.9% in 2025, against 74.5% for ChatGPT. Only 6.5% measure the return on AI against set benchmarks.

A separate GWI survey of 150 British marketers, cited by Digiday, found 29% said inaccurate or misleading AI insights had cost their company money.

On our price sheet, checked 8 November, a Claude Team seat is $25 a month and does not train on your content by default, against $20 for Pro.

With 6.5% measuring AI’s return against benchmarks, the hours that client and agency teams lose to a padded brief show up in almost nobody’s tally of what AI saves.