Amid any confusion and uncertainty as consumer goods companies wrap their heads around the concept, experts say AI will be the only way to meet the demands of consumers going forward.
The supermarket of the future may never run out of your favorite beer – thanks to artificial intelligence-enabled drones. In January, AB InBev announced it’s been trialing the use of autonomous drones and advanced computer vision in a Canadian store to see in real time what’s on shelves, improving product planning and preventing stock outages.
It’s a splashy example of how AI is being used to solve business problems in the CPG industry. But as a use case, it’s far from typical.
Headline makers like these contribute to confusion among CPG marketers about what AI is and is not. In Shopper Marketing magazine’s Trends 2019 survey, only 37% of respondents said their organizations are currently using AI or machine learning (a type of AI). But industry experts we convened for this virtual roundtable discussion said they find that hard to believe. If you’ve built an Alexa skill, you’ve dabbled in AI. Trade promotion optimization software with built-in predictive analytics uses AI.
It could be that folks expect AI to be less like the current state of the art and more like a Steven Spielberg film.
Even if adoption rates are greater than our survey suggests, AI in the CPG space is still in its early stages. But as shoppers expect a more personalized, frictionless, omnichannel experience, the only way to meet their demands is with the aid of AI, our seven experts agree. Here, they discuss how AI is already being used and what to expect in the near future.
What’s the current state of AI use in the industry?
Sandeep Dadlani: I would assume the basic forms of AI are being used at every CPG company to detect patterns and probabilities in data. I would be shocked to find any large CPG companies that are not using machine learning today. It’s commonplace.
Andy Walter: Where folks are starting to have success, they’re applying natural language processing to things like consumer engagement. One of the responsibilities I had while at Procter & Gamble (from 1990-2016) was consumer relations, and for the 6 million consumers every year who contact P&G, our model up until AI was to have a lot of people sitting somewhere answering phones and emails manually. Now, with AI techniques, you can do supervised machine learning and natural voice selection to automate a lot of these processes.
Steven Hornyak: We are at the visionary, early adopter stage, but AI as it relates to CPGs specifically is about to enter what we call the tornado, which occurs when enough people have used it and proved it works so everyone jumps on board.
Where do you see it going in the next couple of years?
Molly Schonthal: We’re evolving from the “measuring and personalizing” phase – an example is the Neutrogena 3-D printed face mask technology; it scans your face and 3-D prints a mask that’s exactly right for you – to “facilitate and simplify,” using behavioral data that is sort of predicting what you might like to see. McDonald’s is using AI and behavioral data to change its digital menu based on who approaches along with other factors, like the weather, so you’re only shown menu choices that are relevant to you in the context of that moment. Next comes “calculating and capturing,” driven by the need for brands to capture consumers and regain control over where and when they show up, be it in an Amazon search bar, be it a smart kiosk, be it your fridge.
Tom Edwards: One, AI will redefine how we as marketers derive insights to inform everything from creative and media, to how to drive client growth. Two, dynamic creative decision engines and the ability to layer and process large amounts of data allow for true personalization at scale. Three, SEO strategy will drastically change as Google and other search providers allow AI-based systems to serve personalized results. Four, multimodal experiences beyond mobile and desktop will emerge as the role of voice and AI-powered virtual assistants create new connection points. Five, as virtual assistants become proxies for consumers, CPGs will need to understand marketing through algorithms that are not driven by emotional connection. Six, the simultaneous advancement of phone hardware and AI creates a pathway for augmented reality at scale. Seven, we’ll increase ROI beyond today’s predictive analytics capabilities using deep learning AI-based systems.
Juliet Noland: Brand storytelling, I believe, is where we’ll see the most impact. Like Pinocchio becoming a real boy – what would your brand look like if it became a living, breathing thing? – which is what AI is going to make it seem like. I think it will create a whole new area of advertising because your brand story comes up against a problem, and that is, when we think about how people are using Alexa or Google Home, they’re automating rote tasks like building a grocery list or ordering laundry detergent, but what’s happening is that AI is then becoming the new gatekeeper of these household decisions. It will be really interesting to see how brands are going to get into a machine’s consideration.
Is it easy for companies to add AI capabilities to existing technologies, or does it require a full-scale adoption of new technologies?
Walter: The technology is more accessible than ever because of cloud-based models and the ability to basically overlay AI capabilities onto your legacy systems. With the right vendor, you can create an orchestration layer or an overlay where AI processes data to solve business problems, and it goes both ways so you then feed recommendations back into your systems for execution.
Rahul Tyagi: One of the big challenges tech companies are tackling is how to democratize AI – how to bring the power of AI to business users so they can just enter their business problem into the computer along with the relevant data to get an AI-generated solution. There are many new technologies in this area, broadly called auto machine learning, that basically hide machine learning’s complexity and expose an easy-to-use interface to users.
What are some of the CPG business challenges AI is currently addressing?
Hornyak: It’s touching finance, it’s touching brand marketing, it’s touching sales, it’s touching supply chain. Ultimately, CPG companies live and die by their sales and revenue. Trade promotion being the largest expense in sales, it’s one of the key focus areas.
Schonthal: A lot of CPGs are using AI to judge content effectiveness, scraping the web using data and information to determine at a much more rapid pace rather than having to wait for results from your analytics company of choice. We’ll see it in marketing automation – it’s already happening, meaning the same experience is rendering differently for me than it is for you – and in digital merchandising and dynamic pricing.
Dadlani: When you use design thinking to find the right problem, AI becomes almost like a super power. At Mars, we use it to drive our growth strategy. An example is our successful launch in Australia of the Snickers Hungerithm, which scoured social websites for thousands of posts every week and adjusted prices of Snickers bars according to the anger levels on the internet. It resulted in fantastic category growth and sales growth, and is a great example of algorithms in action for a particular cause.
In which areas might AI have the greatest impact?
Tyagi: Digital marketing and e-commerce are the areas primed for a significant impact through AI and machine learning technologies because of the volume of data they generate. Digital marketing is basically any online interaction a consumer has with a brand, be it through search, comments posted to a blog or social media site, calls to the brand’s consumer affairs department, etc. When we combine all of this data with e-commerce transaction data, we have what is often referred to as big data. To make sense of and act on such huge and dynamic data in near real time, one needs the power of AI and ML.
Walter: AI will make a big difference in the media space, where spend is so big that even small improvements will have a significant impact.
Noland: In terms of business functions, it will help with ordering and matching supply with consumer demand – making sure you’re stocking the right things at the right time, and creating a more dynamic shelf.
Edwards: Three key words: proxy, prediction, pervasive. Research shows that Gen Z and Millennials are very open to having a virtual assistant serve as their proxy for frequent purchases. AI will accelerate the ability to create predictive models that can optimize spend for brands, and AI will fuel the need for pervasive design. This means designing for more than just desktop and mobile, to also include voice, vision and touch (such as spatial computing).