As the generative-AI craze grips the corporate world, advertising agencies have been eager to demonstrate to their clients — and shareholders — their AI expertise and deep relationships with the likes of Google, Nvidia, OpenAI, and Microsoft.
As agencies guide the world’s biggest companies in adopting AI, they’re beginning to consider how their own businesses too are being disrupted by the technology, including what to automate, where to source AI talent from, and balance potential revenue uplift with business risks.
Some of these challenges, like avoiding copyright infringement issues and disclosing when content is AI-generated, are well-known — but the ad world is also facing a handful of new problems many hadn’t anticipated before.
Here are some of the newest challenges agency leadership teams are discovering as they race to lead AI in advertising.
Some ad agencies fear GPU shortages could delay client work
The rise of companies deploying artificial intelligence has led to a surge in demand for graphic processing units, or GPUs — chips that can process large amounts of data and media simultaneously and at speed. Companies are increasingly being divided into the GPU haves and have-nots.
For now, most agencies with demand for GPUs are simply renting compute power via their cloud service providers like Amazon’s AWS, Microsoft’s Azure, and Google Cloud.
As Big Tech and VC-funded startups like OpenAI look to capture market share with case studies, agencies are reaping the benefits of this competitive environment, said Nick Coronges, global chief technology officer at the ad agency R/GA.
“The benefit for us is that this stuff is relatively cheap,” Coronges said.
But as the market matures, some agencies expect those prices to rise. And as agencies promise clients that they can rapidly generate thousands of creative assets using generative-AI, many will need to start sourcing their own compute.
If they aren’t planning for that now, they could find themselves at the back of the line amid monthslong waitlists, said Lewis Smithingham, svp of innovation at the digital agency Media.Monks. Already Media.Monks is booking GPUs six months in advance for some projects, he said.
“Agencies might find themselves being overcharged for use and purchase, or even more likely, they’ll realize they can’t deliver on their pricing because their in-house compute is out of date, or they underestimated the amount they need to produce the content they have promised,” Smithingham said.
Productivity gains from AI risk undercutting agency fees
Many agency bosses have touted the productivity efficiencies AI presents to their companies — from creating lists of search keywords to making tedious work like cataloging invoices less time-consuming.
But those productivity gains could also end up cutting agency fees.
Agencies typically charge on a “full-time equivalent,” or FTE model, based on the number of hours, equivalent to full-time employees, they spent working on a client’s business.
“With the level of automation now possible through generative AI, it does start to question how sustainable those traditional time and material based compensation models are” said Ryan Kangisser, managing partner for strategy at the consultancy MediaSense.
Analysts at the research firm Forrester have suggested that agencies may need to adapt from an FTE to HTE — “human plus technology equivalents.”
“What is being delivered is not simply human hours, it’s augmented human hours, processing, AI, and all that enriched with a lot of data and expertise,” said Jay Pattisal, vp and principal analyst at Forrester.
Agencies haven’t figured out what to do with the AI talent they’ve hired
Much like heads of digital, heads of mobile, and heads of programmatic before them, agency heads of AI are now having their time in the sun.
But once such AI talent is sourced, agencies are presented with another dilemma: Where in the organization should a “head of AI” sit? Most businesses have started with this being part of the chief technology officer or the IT department’s responsibilities, said Ruben Schreurs, chief strategy officer at the marketing consultancy Ebiquity.
Schreurs and other experts said the AI remit is much broader than technology alone, given the future business impact and the corresponding legal and compliance risks.
As the tech matures, more agencies will move to what Richard Robinson, managing director of the Xeim Engage consultancy describes as “a singular leader armed with the fiscal responsibility and decision-making authority to deliver the necessary recommendations and decisions on agency structure, culture, and purpose.”
Read the full article here