At the opening of the classic song “Folsom Prison Blues,” Johnny Cash sings, “I hear the train a comin’ / It’s rollin’ round the bend.”
After a few weeks of research into what companies are actually doing with generative AI in 2024 — following the trade publications and watching hundreds of press releases roll down the PR tracks — I have the definite sense that the GenAI train has not just left the station, but it is approaching bullet train velocity.
I limited my research to press releases or related media coverage in Q1 of this year. I also focused solely on recognizable brands. As I looked into the backstories of some of these new products and services, I found there had often been internal antecedents, where companies already had AI projects in flight. But it was clearly the ChatGPT lightning strike of November 2022 that caused things to move into high gear.
A Gartner poll from October 2023 suggested that just 10 percent of companies at the time had released GenAI services or features, and another 45 percent were working on pilots. Based on my scan during Q1, I’d reckon that the actual public release needle has been climbing higher since that report.
Below are 17 recent notable projects, broken down by industry.
Automotive
Two major automotive manufacturers, DS (Stellantis) and Volkswagen, have been working on ChatGPT additions to their vehicles’ infotainment systems — the former into the Peugeot “i-cockpit” and the latter into its IDA voice assistant. Both are offering spoken interfaces to drivers like, “what’s the next best place to charge my vehicle and how long will it take to get there?” or “is there any decent FM radio playing classic R&B in this city?” How “colorful” these systems might end up being in their replies, we’ll see.
It’s worth mentioning that both these offerings are paid subscription add-ons. But that’s not the case with Tesla, which just updated its free mobile app. It now sports a chatbot interface called “Tesla Assist Beta,” which can answer questions about Tesla products.
Finally, in an effort to optimize manufacturing, Mercedes Benz (Daimler) is using chatGPT via Microsoft’s Azure OpenAI to help employees analyze production data to improve error identification and quality control.
Banking / Insurance / Accounting
A firm that’s gone all-in on AI is Wells Fargo. As many as 4,000 of the bank’s employees have gone through Stanford’s human-centered AI (HAI) program. This has put dozens of back-office projects into production. The bank’s highest-profile public facing AI is the bank’s virtual assistant, Fargo, which leverages multiple LLMs and Google DialogFlow architecture to let users do things like ask questions about their spending trends in either English or Spanish. CIO Chintan Mehta recently was quoted saying the app handled 20 million interactions since its launch in March 2023.
From the accounting world, a strategic partnership between KPMG UK and tax tech firm Blue J led to a GenAI-powered tool called “Ask Blue J” that can serve up answers to sticky tax questions. The deal gives KPMG UK exclusive access for three years, and the app has live access to Blue J’s secret sauce — a curated database of global tax content.
Online car insurance startup Jerry says it is already saving several million dollars a year by leveraging GenAI to answer customer questions. According to a recent InnoLead case study, the company’s chatbot debuted in May 2023, and by June 2023, 100 percent of inbound messages were getting responses within 24 hours, up from just 54 percent in April 2023. Today, 96 percent of messages are responded to within 30 seconds. The number of queries requiring human intervention has dropped from 100 percent to just 11 percent.
E-commerce & Retail
Auction website eBay just released a tool called Photo-to-Listing to solve what they refer to as the “cold start” problem, which is when a new user has to enter a raft of information to get their first product listed on the marketplace. With this new capability, a user takes a picture of their product and the description is largely written for them.
H&M Group has added a touch of GenAI image generation magic to their existing online Creator Studio for making custom gear. In the past, users created artwork offline and then uploaded it to see their handiwork previewed on a piece of apparel. Now, a magic wand icon with the word “generate” sits next to the upload button. I actually played with this one and produced what I think is quite a nifty t-shirt for innovation nerds (see right). That said, this is certainly a very simple implementation that makes for a splashy headline. In other words, H&M is picking the low-hanging fruit here.
L’Oreal is introducing, in partnership with NTT Data’s conversational AI platform “Eva,” a “virtual beauty advisor” available to consumers through either WhatsApp or Instagram. Trained over months on internal data, the system can analyze a user’s specific situation and then recommend L’Oreal products to match. The hope with this hyper-personalization strategy is to connect with new customers and drive top line numbers.
IKEA, via a customGPT they published to OpenAI’s app store, invites users to upload an image of a room in their home, and then guide the AI with prompts like “show me my future modernist living room with touches of Paul Klee and organic sustainable materials.” The customGPT can recommend specific products from IKEA’s catalog, and offer links out to its site for purchase.
Healthcare & Pharma
In pharmaceuticals, this fascinating interview with Helen Mirianos, Head of R&D Portfolio Strategy for Sanofi, notes that the company aims to use AI at scale. Mirianos details not just the direct impacts to drug discovery cycle times using their AI app “Plai,” but also the transformative aspects it’s having on employees’ decision-making and management processes.
From the world of pharma marketing, the introduction of “Charlie” at Pfizer (named after company founder Charles Pfizer) is all about improving what it calls its “content supply chain.” The app, in use by the internal marketing unit and several of its external ad agency partners, speeds up content creation, editing, and even fact checks and editorial reviews. Longer term, Charlie is being integrated as a just-in-time knowledge resource across their other systems, like Slack, Adobe Experience Manager, and Workfront.
At the other end of the pharma world, in the pharmacy itself, Amazon Pharmacy is adopting GenAI to cut prescription processing times, improve stock accuracy, and add transparency to pricing. For example, on that last point, LLM’s can give real-time insurance estimates on meds without the customer having to enter their insurance data.
For a peek into the future of healthcare, Cedars-Sinai has rolled out an AI-powered immersive app to support patient mental health, using the brand new Apple Vision Pro headset. Dubbed “Xaia,” which is short for “eXtended-Reality Artificially Intelligent Ally,” the experience is described by Dr. Brennan Spiegel as a “quantum leap” in psychotherapy. Patients are immersed in a healing world with an AI avatar that simulates a real therapist. (Also programmed to say “our time is up” at 50 minutes.)
Restaurants and Fast Food
If you’ve driven through Wendy’s in Columbus, Ohio in the past few months, you might have noticed your order taker was exceptionally accurate and speedy in their responses. That’s because the restaurant chain has been testing their so-called “FreshAI” in the Columbus metro area and has shaved 22 seconds off the average order time, according to a company press release.
Similar tests, also being conducted in Columbus by Lee’s Famous Fried Chicken, in partnership with Presto AI, have found similar improvements in order time and accuracy, as well as the most valued metric of all: increased upsells. Incidentally, Lee’s went a step further, licensing the voice of a local football celebrity. Your AI order assistant is not just a bot; it’s a bot impersonating your favorite gridiron star.
Starbucks has had a loyalty rewards program for years to increase customer retention, but their newest upgrade uses AI to hyper-personalize those rewards. Leveraging a massive company data store of analytics known internally as Deep Brew, and a GenAI interface, the chain can identify and incentivize specific cohorts of rewards members. Carol Ruggeri, Starbucks’ CFO, explains the program here.
Lastly, this one is not a well-known brand in the West, but it’s too good of a story not to share. The oldest bakery in Japan, Kimuraya, released a line of new “romance breads” in February. In partnership with NEC Corp., they ran 35,000 love-related song lyrics, as well as conversations from the popular TV show “I Fell in Love with You Today,” through an AI to identify five key romantic emotions or moments in the dating journey. The AI matched ingredients to sentiments and recommended flavors and textures. The outcome is a media sensation and great press for NEC to boot.
Laying Down Tracks
I’ll close with a safe prediction: where GenAI reduces friction, optimizes processes, or improves on any of the classic jobs-to-be-done, it will be adopted at a rapid pace.
…The first movers are laying down tracks that will benefit them in the future.
There will be mishaps, and public embarrassments. Take Google’s recent missteps with its Gemini image generator. C-Suites and corporate boards will fret about liability and risk. But as the risks diminish over time, the compelling profit and productivity boosts will be irresistible.
It’s also clear, after surveying a broad swath of Q1 GenAI announcements, that the first movers are laying down tracks that will benefit them in the future. Like Wells Fargo’s Tachyon, or Pfizer’s Charlie, these players are putting AI data infrastructure in place and running real world testing and deployment cycles, ensuring that their talent will be that much nimbler at adopting what comes next week.
I hear that train a comin’…
