From GTC and GDC to Shoptalk: Major industry events put AI on center stage

From microchips to gaming to e-commerce, three previously disparate worlds all share a similar thread this week: the ever-promising allure of artificial intelligence.

It’s no surprise AI is a major focus at three major conferences taking place this week: Nvidia’s GTC (“GPU Technology Conference”), GDC (“Game Developers Conference”) and the retail-focused Shoptalk. In both Silicon Valley and Las Vegas, countless companies are using the annual events to promote new products and services while riding the current wave of hype. But will more computing and more data lead to more opportunities, or just more noise? As Nvidia heralds the next era of computing, others hope to tap into new ways of using AI for retail, search, marketing and the omniverse. 

“We’ve increased computation by 1,000 times in eight years,” Nvidia CEO Jensen Huang said on Monday during his GTC keynote. “Remember back in the good old days of Moore’s Law, it was 10 times every 5 years, [and] 100 times every 10 years at the middle and the heyday of the PC revolution…The rate at which we’re advancing computing is insane and it’s still not fast enough.”

After debuting its new Blackwell AI chips — named after the renowned mathematician and statistician David Blackwell — Nvidia also announced a number of other updates across its line of products.

New tools include using Nvidia’s Avatar Cloud Engine (ACE) for speech and animation to create “digital humans,” which can interact in retail and gaming environment using natural language processing. The company also announced it will offer its omniverse platform to be streamed through Apple’s Vision Pro headsets, and is working with the Weather Co. to create a digital twin of Earth to understand climate change. Ad agency CTOs that tuned into the presentations said they also see benefits in new tools for optimizing inference across a dozen AI models to speed up training and testing.

Several agencies also used Nvidia’s event to learn about the sector and announce their own news. For example, WPP said it has joined the Alliance for OpenUSD (USD) to help set standards for the future of 3D content across numerous platforms. The goal of USD — which includes other members like Nvidia, Pixar, Apple and Autodesk — is to set uniform standards for 3D formats just like HTML did for 2D.

“Our 3D digital world will be transformed by USD and AI,” WPP wrote in a blog post. “As companies and products become too complex to manage with the screens and pages, organizations will turn to 3D representations of their data. There must be a standard way of saving these 3D files if there is to be collaboration.”

Other companies also announced new AI-powered updates for 3D content. On the same day as Nvidia’s keynote, Adobe announced tools in its Substance 3D content creation software are also now powered by the same Firefly generative AI platform powering products like Photoshop, Premiere Pro and Adobe Express. With Firefly, Substance 3D will add ways to turn text prompts into textures, generate 3D images from a mixture of photos, and add composite images into backgrounds. 

According to François Cottin, senior director of marketing for Adobe Substance 3D, the tools could be used in a range of instances across gaming, virtual reality, e-commerce and automotive. He also likened what’s happening now with 3D images reaching a “tipping point” to what transformed the democratization of video 10 or 15 years ago.

“If you are texturing a violin [product] and you need some wood, you can scrape the internet for pictures and find the right pictures in Adobe stock or wherever you find it,” Cottin said. “You can take your own picture, but you can now also generate many variations of wood until you find the exact one you want.”

One of the main e-commerce barriers has been generating net-new creative assets, said Ian Mackenzie, chief creative officer at McCann Worldwide Canada. Rather than figuring out how to get AI models to generate perfect cars, phones or other products, generating backgrounds and 3D models for various products could be helpful when imagining different colors or different angles.

“On first look, maybe Firefly is good at switching out backgrounds and generative elements in an environment that helps merchandise products in more personalized ways,” Mackenzie said. “Then we can use pretty traditional assets and build generative worlds around them in a way that drives e-commerce activity without putting the full pressure on a generative model to be generating products.”

How generative AI is powering new types of text

Other companies are adding new ways for retail marketers and others to use AI when generating campaigns for email, text messages and entire websites. During this week’s Shoptalk in Las Vegas, conversational platforms previewed new AI-powered features. Persado added ways to use its large language model for optimizing website copy, SEO and audience segments. Another, the SMS-focused startup Attentive, said it’s now letting retail brands personalize opt-in text messages to market products and other marketing based on website visits, abandoned carts and shopping habits.

Although branded SMS marketing has so far been mostly one-way, Attentive CEO Amit Jhawar said consumers are starting to want more real-time responses. AI models like ChatGPT and Gemini are helping consumers to better understand how to interact. Jhawar, who was previously CEO of Venmo, also noted that it helps when retailers prompt consumers questions like they might in a physical store, like “What brought you into the store today?” or “Any questions about this item?”

“Another learning from Venmo is one of the biggest reasons consumers abandoned shopping carts after they’ve added something to cart,” Jhawar told Digiday. “They have a lingering doubt about the item and they can’t get the last [question] like ‘Does this sweater pill?’ or ‘Does this shirt have a waterproof texture,’ or ‘Is it wrinkly or not?’”

Meanwhile, retail giants have been adding new AI tools for e-commerce companies and shoppers long before this year’s Shoptalk. In January, Walmart added new generative AI features for searching on the retailer’s website. Last week, Amazon added a new way to help sellers generate detailed product listings based on a URL. And earlier this week, a Google exec told Glossy that its new image generation tool called “Dreamer” is poised to change how people search for outfits online.

Infusing AI into e-commerce also has potential pitfalls. Along with concerns about fake products potentially scamming shoppers, other experts worry about how fake AI-generated product reviews could clutter platforms like Amazon, prompting the Everything Store to call for both the private- and public-sector partnerships to help address the problem before it scales. Another concern is AI-generated books that compete with human-written books, potentially creating an additional headache for authors already fighting with AI companies over copyright concerns.

Agencies and startups are also developing new tools to help personalize e-commerce pages. A startup called Lily AI just raised $20 million to expand its e-commerce product discovery platform. Another startup, Algolia, also recently added new features for AI search. Meanwhile, the media buying and creative agency Dept has built AI features that personalize websites and more naturally integrate generative AI into a brand’s existing user interface.

“One of the areas we think is the least appreciated and most overlooked — and we still can’t figure out why — is the completely transformative impact that generative AI will have on customer experiences,” said Isabel Perry, who leads emerging tech at Dept. “What mobile did to responsive design, this is an attempt to be a new version of generative UX, and so we’re building these brand language models to basically integrate with when you’re looking at commerce experiences.”

Retail-focused AI isn’t just for e-commerce; it’s showing up in brick and mortar as well. Companies like CoolerX — which recently rebranded from Cooler Screens — aims to offer more personalized and contextualized marketing to shoppers. As shoppers peruse fridges and shelves, the startup uses sells ad space on refrigerator doors powered by predictive models as an in-store digital retail media network. The goal, according to CoolerX chief growth officer Anand Muralidaran, is to provide a “physical equivalence of Google or Facebook or Amazon.”

“I’ve always believed that stores are a critical part of a journey for a consumer,” said Muralidaran, who previously led Nvidia’s AI and machine learning division for retailers. “And a store is a critical part of a channel because nobody wakes up in the morning and says ‘I’m going to wake up and shop at Target, but only through the mobile app or website.’”

https://digiday.com/?p=538590

More in Media

AI Briefing: Senators propose new regulations for privacy, transparency and copyright protections

A new bill called the COPIED Act aims to pass new transparency standards to protect IP and guard against AI-generated misinformation.

Media Briefing: Publishers reflect on ad revenue midway through 2024 

Some publishers say ad revenue is pacing 15% up year over year while others are still managing their expectations for how 2024 will shake out.

Teads is exploring sale options as M&A in ad tech heats up

Sources state the Altice-owned stalwart of outstream video has recently held talks with private equity and strategic players.