How NVIDIA Is Unlocking the Power of AI for Advertisers
Jamie Allan spent his formative years as a photographer and director, before shifting into technical consultancy roles. After a decade of designing and implementing solutions for post-production, broadcast and VFX facilities, he moved to NVIDIA where, after six years, he now leads its efforts with global advertising agency groups and building partnerships in that space.
The director of business development, global agencies and advertising describes NVIDIA as a “very vertically focused organisation”. The company has developed foundational layers of technology, including hardware and a comprehensive software stack that they then take to market with specific vertical solutions. For years, NVIDIA has helped the media and entertainment industry solve problems, especially with 3D content creation and real-time rendering - some of the fundamental use cases for the GPU that they are proponents of.
Over time, they have ventured into new areas such as video broadcast pipelines, providing advanced technology stacks for ‘computer vision’ - enabling the development of Hawkeye technology, which has been a mainstay of tennis, cricket, baseball and other sports broadcasting to date.
But now, AI is the talk of the town, and NVIDIA is right there at the epicentre of its emergence - especially in ad land. LBB’s Ben Conway caught up with Jamie to find out more.
“In the advertising and marketing space, there has been a growth curve like no other with AI in the past few months,” he says. He shares that AI usage in the media industries has been accelerating for the last five or six years, becoming more commonplace not just in broadcasting, but also in content creation and VFX - with companies like Adobe and DaVinci Resolve incorporating it into their software. New AI-first companies have also been leveraging AI to create new tools, such as AI content innovator, Metaphysic, which says it uses “NVIDIA powered super-computers” to run its proprietary software that can de-age and create synthetic humans for TV and movies.
“What's happened in the last six, seven months is the ‘iPhone moment’ of AI,” he continues. “Chat GPT and everything around November last year put the power of AI at the front of mind, both for brands’ CMOs and CEOs, as well as for every agency leader and every creative leader in this space. And it's just accelerated the interest of working with NVIDIA.”
After working with media companies and advertising agency groups for some time, Jamie and his team developed a close relationship with WPP and its SVP for creative AI, Perry Nightingale. Together, they decided to work on evolving the content supply chain for marketing, using AI to scale and push new solutions forward.
“We build foundational technology that developers and organisations then build on top of, and I think the work with WPP has been a great example of that,” says Jamie. “They've taken our Omniverse platform for creating and developing 3D worlds and complex 3D content solutions, and built a solution on top of that which they're now taking to market with their clients.”
One of the more prominent places this Omniverse and AI-powered work will appear for WPP, according to Jamie, is in its automotive work. This is largely because the Omniverse platform allows the laborious tasks involved in using CAD data (design data for geometric models) to be accelerated and, in some cases, automated - while still allowing the creatives and engineers to use the same tools they’re familiar with.
Omniverse - and its cloud-based version, Omniverse Cloud - flow into NVIDIA’s ‘graphics delivery network’, a new platform that delivers real-time 3D experiences to consumers across web, mobile and tablet. “This is where we’re expecting the fruits of our labour and collaborations to be realised,” says Jamie, adding that when combined with “indemnified, commercially viable licensed generative AI tools” - such as the one NVIDIA is working on with WPP and Getty - creatives will be able to produce more variations of 3D imagery, more quickly and easily.
The feedback from brands and clients has shown NVIDIA that this is a sought-after capability, as Jamie explains, “Variations on particular content for personalisation drives more commercial growth for product sales. So being able to create more variations easily and quickly with that combination of a connected 3D ecosystem, powered by NVIDIA Omniverse, and all of our existing partner tools from Adobe, Getty, Shutterstock, etc. creates a platform and a toolset that will be incredibly powerful for the industry.”
As well as working with WPP for the last several years, NVIDIA also has existing long-term relationships in the advertising space with the likes of Google, Meta and Adobe. For Jamie, it’s this network of collaborations with brands and agencies that will empower creatives and allow them to build new tools and capabilities for their needs. “They're going to be supercharged with tools like NVIDIA Omniverse, with platforms and capabilities from Google and Adobe, with NVIDIA’s AI enterprise platforms and with other partners that we’re growing.”
Highlighting partners like Israeli generative AI company BRIA, which is pioneering image tooling and pipelines, as well as Polish deep-tech scaleup SentiOne, which applies AI to audience insights, he adds, “It's really about this much bigger ecosystem than just leveraging tools directly and platforms from NVIDIA. We’re enabling every end of this ecosystem and story”
The challenge, however, for many companies - especially larger, well-established ones that lack the agility to adapt efficiently - is finding where in their creative process and structure to fit AI (and other new tech in general). Jamie explains that “doing AI is a scale problem” and that, to be successful, AI must be centralised so it can “affect and provide services and capabilities to the whole organisation”.
Other challenges stem from AI’s innate necessities - being a naturally computing-intensive and engineering-intensive process. Forming a team that understands the technology and knows how to build AI well and safely is fundamental, says Jamie, explaining how leveraging AI at scale is reliant upon safety and guardrails. For this exact purpose, NVIDIA has developed tools like its NeMo Guardrails - an open-source framework to enable developers to easily safeguard their AI for chatbots and image generation. He says, “The legal and compliance teams at every brand and every agency group need to have as clear of a strategy as the actual creatives and the AI teams do.”
Supporting the efforts to regulate AI, especially when used for consumer interaction, Jamie adds that NVIDIA works with government groups, the Content Authenticity Initiative and even DARPA, the R&D agency of the US Department of Defense. The tech company also works with individual agency groups and brands within adland to build and maintain brand-safe and brand-compliant AI models - which require fine-tuning and a deep understanding of their data.
But if a company can achieve this, he says, advertisers’ work won’t change fundamentally. Creatives will still be exactly that - creative. They will still generate and curate the ideas, however, he suggests that the speed at which an idea goes from brain to canvas will be accelerated. This reduced time comes from AI tools like Omniverse and harnessing the power of USD (Universal Scene Description), a format for 3D CG scene descriptions that contains data for the geometry and appearance of 3D models, animations and virtual cameras.
“It's the ability to interchange 3D between many applications. That's what allows us to power the collaboration and streamlining of the content creation process between many applications.” He continues, “It's what's powering digital twins at companies like BMW and Mercedes, it's what's powering robotics development and our autonomous driving platform. It was originally created by Pixar to accelerate the work in 3D and VFX. Now, we're bringing that power through Omniverse to many industries.
“Not only does USD allow you to collaborate between multiple teams, but it becomes very easy to integrate AI tools, which could be coming in from a CRM system and recommending how you put a virtual world together or how you create something personalised for a client.”
While the ability to generate 3D environments and personalised assets at speed will take the headlines, Jamie says that AI has a more under-the-radar use case in the world of advertising today. Data science and machine learning have long been employed to navigate the bounty of first-party data many advertisers have at their disposal - but thanks to AI, predicts Jamie, this process has significant untapped potential that “may have a bigger impact”, despite being “less sexy” than its other uses.
“It will become incredibly powerful in the application of large language models to data insight. Being able to interrogate your data through conversation… or being able to give content to a language model that understands a particular audience is really interesting.” He adds, “Some of these other ideas will help inform what content we should be creating, which makes the process more efficient and have more impact, rather than just creating for creation sake.”
