Agencies say AI is making marketing briefs longer and harder to act on

Agency leaders say generative AI is making client marketing briefs longer and harder to act on, not faster or better to produce, according to comments from agency executives at Digiday's Media Buying Summit this week, reported by Digiday on November 5.

What changed

Nadja Bellan-White, North American CEO of M&C Saatchi, told the summit that AI use in brief-writing is nearing a "tipping point" and estimated that roughly 85% of the client briefs she now sees arrive already written by AI. She argued the tool still has a place, but only as a sparring partner while a human keeps the final judgment calls.

Other agency executives, some speaking under Chatham House rules, described a familiar pattern: briefs that read as disjointed, padded with jargon, and missing a clear through-line to a single marketing objective. Tony Russo, VP of innovation and strategy at independent agency Mile Marker, put it more bluntly, calling the pattern "a judgment problem," not a tooling one. Christina Riccitelli, chief innovation and analytics officer at media agency OpAD, said the fix has to come from management: senior staff need to step in when junior teams lean on AI output instead of thinking a brief through themselves.

The scale of adoption backs up what they're describing. A Basis Technologies survey fielded in July and August found that 56.4% of marketers and agency staff now use AI to help build strategy, media or creative briefs. Digiday's report also cites a 2025 MIT Media Lab study that found measurably reduced writing-related brain activity among people who drafted with AI assistance compared with those who wrote unaided — one proposed explanation for why AI-assisted documents read as competent but hollow.

Who it affects

This isn't limited to agencies. A separate, first-party survey from research firm GWI, published September 30, found that 72% of more than 450 marketers across the UK, US and Singapore said their organization had made a significant business decision based on an AI-generated insight that later turned out wrong or misleading. Among the consequences: 37% reported financial losses tied to those decisions, 44% said inaccurate AI output forced teams to redo work, and 41% said it delayed decision-making. Separately, 91% of the same group said the risk of basing decisions on bad AI output is significant — yet only 37% said they actually verify AI-generated material most of the time.

"AI is only as reliable as the data behind it." — Jason Mander, Chief Insight Officer, GWI

For a software marketing team, a bloated or jargon-padded brief isn't a style problem. It's the document that sets media budgets, creative direction and success metrics for a campaign. A brief with no clear objective produces creative with no clear objective, and an agency team under deadline pressure has every incentive to let AI fill in gaps the client never resolved.

B2B software marketers are exposed on both ends of this. Internally, teams are using AI the same way agencies are — to draft positioning briefs, competitive summaries and campaign plans faster than a person could research them from scratch. Externally, most SaaS companies brief at least one outside agency or freelance team for paid media, creative or content, and that brief is now more likely than not to have passed through an AI tool before a strategist ever reviewed it. Neither side of that relationship has caught up with a review process built for AI-assisted drafts rather than human-written ones.

What to do differently

None of this argues for banning AI from the briefing process — adoption is already too broad for that, and the agencies quoted aren't proposing it. The shift is in where human review sits in the workflow.

• Require every brief to state one marketing objective in a single sentence before any AI drafting happens, so there's a fixed point to edit back toward.
• Treat AI-drafted sections as a first pass that a strategist condenses, not a final document that gets forwarded as-is.
• Ask your agency directly what share of the brief was AI-drafted and who reviewed it — Bellan-White's 85% estimate suggests the honest answer is often "most of it."
• Spot-check AI-sourced data points in briefs and decks against a primary source before they inform budget or targeting decisions, given how many marketers in the GWI survey said they skip that step.

The deeper risk agency leaders are flagging is cognitive offloading — handing the thinking, not just the drafting, to a tool. A long, well-formatted brief can pass a skim review while carrying none of the judgment a shorter, rougher one would have forced someone to apply. For software marketers managing agency relationships on tight budgets, that's the costlier failure mode: not an AI-written brief, but an AI-written brief nobody actually re-thought.