Agencies find themselves in something of a bind when it comes to AI. On the one hand many agency executives, considering the considerable competitive pressure of the agency market, believe it’s vital that they adopt AI faster and more effectively than their competitors, in order to build or maintain a competitive edge. Some of the major groups are already reporting that their AI capabilities — be it content creation or AI-driven targeting and measurement — are helping them win pitches.
At the same time, clients see potential risks in how their agency partners use AI. Research from the World Federation of Advertisers (WFA) last year found that eighty percent of brands surveyed expressed concerns about how their agencies use AI, citing ethical and reputational risks.
Part of the solution lies in providing clear terms governing agencies’ use of AI in their contracts with clients. In the WFA’s survey, just 38 percent of brands said they had AI related terms in their contracts. UK advertiser trade group ISBA is trying to bump that number up, introducing generative AI terms into its latest Media Services Framework (which is essentially a template advertisers can use for creating agency contracts).
Managing risk
Writing these sorts of terms isn’t straightforward, given the tension described above. Brands want to protect themselves from risk, but if contracts are too restrictive, any potential efficiency gains from AI could be lost.
Nick Louisson, ISBA’s director of agency services, says he’s seen contracts at both ends of the spectrum.
“We heard of an example where an agency was putting terms in front of an advertiser that said, if we use generative AI, all the risk sits with you,” said Louisson. “So we will notify you that we’re using it, but if anything goes wrong, that’s on the advertiser. That terrified me!”
In reality, this sort of contract will block any use of generative AI anyway, since no brand legal team is going to agree to those terms.
At the other extreme, Louisson said he’d seen a document where “all uses of AI needed to be approved by the advertiser, but it didn’t constrain that to generative AI, it was all AI […] and that raises questions — my laptop is filled with AI, so many apps use AI.”
ISBA’s aim therefore has been to create a framework through which advertisers and agencies can communicate clearly about how and when AI can be used, in a way which encourages (rather than discourages) AI adoption.
For a start, the new terms relate specifically to generative AI, rather than all AI. The cat’s already out the bag with AI more broadly, as it’s an ingrained part of media trading, and it’s generative AI specifically which advertisers are grappling with.
Beyond that, ISBA’s approach has been to think about the main variables when it comes to risk with AI. For example, legal risks vary significantly based on which LLMs an agency uses, and whether those LLMs have licensed the underlying data which they’re trained on.
“So one of the things we’re asking is what platforms are we using? What safeguards are we putting in place? What records do we need to keep, so if something does go wrong, we can actually identify what happened and demonstrate that we were responsible?” said Louisson. “Pretty much everything that we put in [the framework] was to incentivise an agency and an advertiser to talk to each other about the steps that they can take so they’re comfortable using generative AI. So what are the use cases an agency would use, what are the tools they would use, and how would they use them?”
Even with these terms in place, the speed with which AI tools evolve and add new capabilities complicates matters. But Louisson said that for the time being at least, big advertisers are generally cautious when it comes to AI. Thus they’d prefer terms which restrict agencies from using new AI capabilities as soon as they come out, but which are safer from a risk perspective.
Setting principles for principal media
Another big change to the Media Service Framework is an update to terms relating to principal media (also called proprietary media and inventory media).
Previous iterations of the framework had already introduced terms covering principal media, and Louisson said these had helped improve transparency for advertisers. But consultations with members ahead of the recent update had indicated there was room for further improvement.
There’s an inherent level of in-transparency with principal media. ISBA understands, for example, that agencies aren’t going to show the price they paid for media which they then sell on to clients through principal media deals.
But Louisson said there’s significant variation in the levels of transparency offered by different agencies. The aim for the framework, therefore, is to ensure that advertisers understand what they’re signing up for, and to set guidelines for agencies to work within.
“We want advertisers to create a framework that is very clear and transparent in what they’re getting, in how they want to manage it, in who needs to be approving it at what time, and that they contract that with their agenc so that the agency works within that framework,” said Louisson. “That means the advertiser is confident that their expectations are being met, that they’re getting the level of transparency that they expect, and that [principal media] is being bought essentially by the decision trees that advertisers have set up for determining whether or not it is appropriate in a particular scenario.”
A contentious area
ISBA’s update has proven contentious. The Institute of Practitioners in Advertising (IPA), a trade group representing agencies, released a statement expressing concerns over the new framework, accusing ISBA of touting an “anti-agency narrative”.
“Nick [Louisson] seems to be suggesting that all media agencies are engaged in some form of systemic malpractice around inventory media,” said Richard Lindsay, the IPA’s director of legal and public affairs, in a statement.
Louisson told VideoWeek that one of the reasons for updating the framework’s terms around principal media was due to compliance concerns, which had arisen in advertisers’ contract compliance audits. But he added that while it’s hard to know the real cause, his sense is that the problem likely comes down to agencies and advertisers interpreting contractual clauses differently (rather than agencies intentionally going against advertisers’ wishes). So the updated framework seeks to add more clarity, and avoid these divergent interpretations.
Clarity won’t fix everything. The IPA’s Richard Lindsay suggested that ISBA’s framework is too complex, making it unmanageable for agencies. “It is not surprising if agencies were unable to comply with each and every one of its provisions,” said Lindsay. “How could they?”
Louisson acknowledged that principal media is a tricky area, again because of the differences in how agency groups operate. But he maintained that for advertisers, it’s key that they know exactly what they’re getting into, and retain ultimate control in any principal media agreements.
“It’s difficult to have a one size fits all contract,” said Louisson, “so the ambition would be for advertisers and agencies to just get into the detail, talk about it, and amend those terms to accurately reflect what the advertiser’s expectations are and how it presents itself with the agency, with transparency and appropriate governance in mind. That means that the right people have the right information at the right time to approve or to make decisions as to when the advertiser wants to engage in the practice or not.”
Follow VideoWeek on Twitter and LinkedIn.
