What you need to know:

• WPP production unit Hogarth is working on AI-generated ads for L’Oreal, one of the world’s largest advertisers.
• L’Oreal’s ANZ procurement chief, Orrie Tzamouranis, thinks the technology is already viable – now both brand and agency need to hone human processes to get the speed, scale and efficiency he craves.
• While procurement is often maligned as purely price driven by media agency bosses, Tzamouranis didn’t rule out paying more for ads that prove more effective.
• Either way, Hogarth Future Imaging Manager, Rafael Ruz, said “cost efficiencies … become more dramatic as we scale up the number of outputs required”.
• Slides shown by WPP at SWSX suggest the firm is working on similar projects with Big W, Coca-Cola, Country Road, Dyson, Ford, Kraft Heinz, Mondelez, Nestlé and Woolworths.
• WPP global AI chief Daniel Hulme previously told Mi3 the plan is to build “brand brains” for WPP’s major advertisers and plug those brand asset repositories into generative AI systems in order to build oven-ready creative that doesn’t require endless prompting.

WPP production unit Hogarth is working on AI-generated ads for L’Oreal, one of the world’s largest advertisers. Its ANZ procurement director is “very happy with the outcome” as it bids for scale and speed of production across 34 brands with content that can be localised for key markets around the world.

The ad holding company’s global AI chief Daniel Hulme in March told Mi3 that so-called “brand brains” – repositories of brand assets “glued” to generative AI tools like ChatGTP in order to make oven-ready campaign assets based specifically on brand data – were just months away.

That echoed earlier comments by Hogarth’s local boss Justin Ricketts that AI-powered automated production is the only way that brands can keep up with the explosion in content requirements – with already stretched production pushed to breaking point by the runaway acceleration of retailer media.

Hulme said if the first ‘brains’ could be made to fully fire, WPP could ultimately roll out the technology across its client portfolio.

The aim is to build “mini brains that are trained on a client’s data set, then gluing that brain with a ChatGPT brain,” according to Hulme. “So when a creative says come up with a new campaign that involves a laptop, it knows you are talking about Dell. It uses Dell’s tone of voice, style guide etcetera. So you don’t have to use [additional] prompts to build that context,” he told Mi3. “So the future is gluing together these specialised brains.”

Brand brains

Hogarth has lifted the lid on one of its prototypes for L’Oreal. Per a slide showed at SXSW last month, it appears to be building similar ‘brain’ projects for Big W, Coca-Cola, Country Road, Dyson, Ford, Kraft Heinz, Mondelez, Nestlé and Woolworths.

Hogarth Future Imaging Manager, Rafael Ruz, said the firm has built hybrid AI ads for L’Oreal, for example using the machines to generate backgrounds within ads for brands including Lancomb’s Trésor perfume, enabling it “to extend backgrounds and add elements on the fly very quickly.” He said the ads produced could also be “upscaled” for print use.

In Australia, Hogarth has been working on AI-generated ad assets for L’Oreal’s Essie brand, creating concept backgrounds for the 34 different tones of its nail polish, and ultimately ads that could be used within digital channels.

While Ruz said the overall process remains “time consuming” with “far fewer of us on this on this project” than a traditional production brief, plus tools that are effectively evolving in real-time, “we’re able to ideate some of those concepts very quickly in hours and get feedback from the client”. He said “obvious cost efficiencies … become more dramatic as we scale up the number of outputs required”.

AI procurement

L’Oreal’s ANZ procurement director, Orrie Tzamouranis, said the firm is “very happy with the outcome”. He sees potentially major economies of scale.

“Depending on the year, we are either the third of the fifth largest advertiser by media spend globally, said Tzamouranis with 30-plus brands housed within four divisions across consumer mass market, luxury, professional and dermatological beauty. Given that range, “and the amount we spend on media, we have both very big breadth and depth of requirements for content”, he added. Which is what prompted the firm to work out how AI could do the heavy lifting.

The procurement department is looking for volume at pace with content that can be localised and individualised, per Tzamouranis. The end product must be “as good if not better” than traditional production.

“Scale and speed, which are closely linked, are probably what I'm most interested in from an operational efficiency perspective,” said Tzamouranis. “We still have a little bit of work to do there … But it's quite obvious that as we keep doing this more and more, define the process, rinse and repeat, that won't be an issue in the short-term.”

He said the firm is now getting to grips with how to brief a generative AI campaign versus a traditional brief: “We actually didn't think that would be any different, but it is a bit different.” With “200-plus images” coming back from its first concepting round on the Essie work, the brand team also had to nail “really defining what they wanted, which was actually more challenging”.

Tzamouranis indicated that L’Oreal’s procurement arm sees significant efficiencies incoming in short order.

“I think we have some work to do on the operational side, which is not a huge amount we can't overcome. But we're very happy with the outcome – and I'm interested to see where we are in six or 12 months.”

Asked by Mi3 about the tradeoff between cost and effectiveness, with procurement departments tending to be viewed as primarily cost-focused, Tzamouranis said the firm would have to evaluate how the AI creative performed once launched into real-world media channels. However, he didn't rule out actually paying more: "If there's a brand that maybe has a higher profitability, they're probably willing to expend a bit more on the content."

But he reiterated that L'Oreal sees the AI-driven assets "being entirely viable in the short-to-medium term" with "ways of working and the human element" the immediate hold-ups. "But that is an area for all of us [to work on]. As far as the technology's capability [it's there] ... So I think it is just how do we improve the ways of working to support this new technology?"

Hogarth’s SXSW slides suggest it is firing ‘brand brains’ with systems from Adobe, Epic Games, Getty Images, Google, Meta, Microsoft, Nvidia, Open AI, Runway, Shutterstock and Stability AI.