The development of full artificial intelligence could spell the end of the human race. It would take off on its own, and re-design itself at an ever-increasing rate. Humans, who are limited by slow biological evolution, couldn’t compete, and would be superseded… Stephen Hawking, BBC News, 2014

I think it is assumed that artificial intelligence is here to stay. It is, of course, taking over the world, they say. Or is it?

Do people see artificial intelligence as hype? Does AI fit somewhere on a curve? Does Taylor Swift? Are some things ‘beyond’ the curve?

Blockchain in recent years was touted as the next great thing. Where are all those blockchain use cases, such as food and fair-trade tracking, provenance and others? I just, personally, do not hear about it anymore.

I set out to dig into if AI is in fact hype. I embarked on a set of interviews with a handful of retailers and did some accompanying research to get underneath this. And maybe more importantly, I sought to gather information on real-world use cases being utilized today.

What Is AI?

I think by now, most of us know what artificial intelligence is, but so we can get on the same page, artificial intelligence was a term coined by emeritus Stanford Professor John McCarthy in 1955, defined by him as “the science and engineering of making intelligent machines”; today we emphasize machines that can learn, at least somewhat like human beings do. Generative AI creates new content and data, while more traditional AI solves specific tasks with predefined rules, as per Google. Most of what I will be referring to in the following will refer to generative AI.

Hype – Or Not?

I interviewed a handful of retailers pertaining to real-world use cases – no theories, use cases actually in place today, in addition to doing research myself; subjects included the Chair for a major retail AI council (and currently a specialty retailer Chief Operating Officer), a Vice President of IT with a specialty and department store background, a seasoned Chief Information Officer and a technical Fellow at one of the largest big box retailers in the world.

Going into this project, I did not believe that anyone I spoke with would think that AI had already had the wind taken out of its sails! To the contrary, I had wondered if AI lay ‘beyond’ the curve and if the curve for AI may extend into perpetuity.

The Fellow that I spoke with thought the opposite. He felt that AI is already on the downside of the hype curve and that the hype and buzz around AI would crater in the next year. His rationale: AI has been overinvested in; and its content has to be ‘primed’ by a human. He also asserts that companies may ‘gain’ on workforce reduction, but that those gains will not be offset with commensurate productivity gains. Let us call his view the skeptical vantage point.

On the flip side of this perspective, with a boundless sense of optimism for the subject, the AI retail chair that I spoke with feels that AI is “revolutionary” and that it continues to improve and will just get better with time. He feels that AI will “evolve to justify the hype”.

This same individual gets asked frequently what his ‘AI strategy’ is. He has none, he says. What he does have are “business initiatives of which AI is a part”, where it makes sense, only. We may not know where exactly AI sits on the hype curve today, until sometime in the future, when we look back on its successes, failures and adoption rates.

Let’s have a look at the real-world use cases that retailers are actually doing today.

Sampling Of Real-World Retail AI Use Cases, By Category

Software Development

All the individuals I interviewed told me they are using AI to facilitate software development. None of them just let the AI run on its own. Everything was framed up by a human and checked once code was produced. Java-based development was one such use case mentioned. One of the examples provided pertained to Amazon Bedrock, which according to Amazon is a “fully managed service that makes high-performing foundation models from leading AI startups and Amazon available for clients’ use through a unified API.” Widgets are available and produce code for developers to review, cutting human involvement and FTE time. Humans do a quick check and look-through before go-live. The AI helps with syntax; development staff handles key logic.

The CIO that I spoke with uses GitHub Copilot. His senior developers use it to lay out the framework and the logic, and the AI takes a stab at the foundation, as laid out by the developers. As above, once the code has been produced by the engine, humans review and make changes, as necessary.

Customer Service And Call Center

The holy grail use case from my research appears to be call center applications. This may be obvious and I think many know this may be the place to start, but I have highlighted here a few use cases and companies that appear to be doing it ‘right’.

Some may have heard that Klarna, the Swedish global payments and shopping service company, entered into a strategic partnership with OpenAI, utilizing their ChatGPT technology into a plug-in for shopping. Klarna is one of the first anywhere to integrate it.

The Klarna AI assistant handled two-thirds of customer service chats in its first month, which was January 2024 (and 2.3M conversations, so far). According to Fast Company, the natural-language interface initially helped customers choose items and make other shopping-related decisions, based on personalized queries, a feature Klarna has described as ‘smooth shopping’. Klarna had laid off 700 people in 2022. The company boasted, according to Fast Company, that the AI assistant is doing the work of 700 full-time agents.

The capability may need a touch of time to mature and grow, though. I used it myself and have never used Klarna before. I asked the bot to tell me about my account and the bot replied saying I had account history, which I had not.

Klarna is anticipating that the chatbot could help improve its profits by $40M in 2024, according to Fast Company.

Alibaba Group, China’s largest eCommerce company, has nearly one billion annual active Chinese consumers who make hundreds of millions of transactions daily using its Taobao eCommerce platform. Alibaba uses five AI-bots for customer service engagements, these include bots for merchants, customers, even a simulated customer bot.

The company’s ‘Alime’ bot helps the end-user consumer(s). It is used in online and phone channels and relies on a rich set of interactive UI components that can provide text dialogues, cards, graphics, videos and other conversational interactions between robots and consumers, according to AI Business. This bot also possesses rich voice capabilities and a voice UI to serve customers who prefer to utilize phone channels.

The Alime process can proactively engage with end-user customers and can act as intermediary during service disputes, helping to minimize staff involvement, and just use humans to interface, where needed, as an example, in complex situations or escalations.

Hyper-Personalization

The holy grail of personalization is hyper-personalization and one-to-one marketing. And there is a balance to be had with this, not enough personalization and the ads are just ho-hum and often ignored; too much personalization can be creepy. Target, famously, knows this all too well, and this is a 12-year-old story!

Some are rolling up their sleeves in regards to one-to-one, and trying to do it right; sharing here a few of those stories.

Kroger acquired retail data science firm, 84.51°, in 2015. This acquisition is credited with helping Kroger deepen its omnichannel capabilities and presence. As a means to deepen its ‘engagement’ with customers, rather than just ‘sell, sell, sell’, 84.51° and Kroger are beginning to emphasize personalized nutritional information and recommendations that may boost customer engagement over time. The retailer’s OptUP program uses Kroger loyalty card data to calculate a nutrition score from a customer’s recent purchases. Shoppers can also browse an app while shopping to see nutritional scores of individual products, and receive “Better for You” recommendations for healthier product options.

One-to-one models can be difficult for companies to build that do not have the requisite data. Third-party aggregators are increasingly packaging and drawing data from multiple sources to provide personalization attributes, to sell to retailers and other markets. They might combine for example, customer’s web-browsing history with credit card purchases, social media activity and other characteristics and buying signal data, according to Management and Business Review.

Another example of ‘smart’ one-to-one marketing is Starbucks AI Platform, Deep Brew. Deep Brew was Starbucks’ first foray into machine learning.

The Deep Brew model “optimized the criteria of total revenue from a sale and the likelihood of the customer buying additional items, beyond their normal purchase”, says the MBR Journal. The goal was to personalize across all touchpoints and channels. Starbucks now uses these models to create more than 10 billion hyper-personalized recommendations a year. And the models learn and evolve continuously with new data inputs. Recommendations and data capture include such things as vegetarian food, price sensitivity, tea vs. coffee, etc. Recommendations also personalize the drive-thru experience as well, as an example, on a day with a particularly long drive-thru lane, the order board may suggest quick-to-make items, to help facilitate faster movement of the lane.

Worker Productivity And Engagement

Worker productivity is another potential area of ‘low-hanging’ fruit for the meaningful use of AI. As above, the focus being on identifying real business needs and the potential use of AI to facilitate these, rather than the other way around. The COO that I spoke with saw deep benefit with worker productivity use cases.

This executive has a deaf teammate. Microsoft Teams can convert text, as in virtual meetings, real-time, for the benefit of his teammates who are hard of hearing. Also, his employees use AI to translate from English to Spanish in real-time. As an example, they have a lot of Spanish-speaking employees in their distribution center; they use an AI engine to take speech and text and translate real-time to Spanish for the benefit of these Spanish-speaking employees. And the same for training manuals, AI can be used to translate to Spanish and other languages for employees who speak languages other than English.

Also, on virtual calls – his employees in France, as an example, can listen in to web conferences in their native tongue, with AI translating English to French, in real-time.

eCommerce And Merchant Marketing Support

The final use case area I will cover here is eCommerce and merchant support. Alibaba Group is testing generative AI tools for Taobao and Tmall merchants with the goal to drive sales at its retail properties; these tools are provided via a website called Huiwa.

These tools enable merchant users to create both text and graphics content, providing support to merchants who would otherwise have to do the information gathering themselves, thereby saving them time, and in saving them time, saving them money. A merchant, for example, can input a brief description of a product, which prompts the tool to create customized content for marketing and promotion. The tools are also capable of producing images based on merchant specs and needs.

In Conclusion

Is AI hype? Probably. Is it going away? Absolutely not. Will it ‘crater’ next year? Probably not, but the ‘hypeness’ around it may begin to ‘smooth’ out. As my chief operating officer friend shared with me, the use cases will only continue to evolve and deepen over time. Will AI be the end of all mankind, as Stephen Hawking had predicted? Only time will tell and that story is a topic of another article. Of the use cases that I shared, anecdotally at least, software development seemed to be the furthest along, in terms of being in use, versus being in beta or testing stages. Another theme echoed from everyone I spoke with: human involvement is needed to validate AI outputs. No one is using AI content or outputs without human intervention for checks and balances. Finally, I heard that, really, AI is currently in the stages of exploration and discovery with basic use cases being used presently. Maybe the ‘sexy’ stuff will come later.

Best of luck along your AI journey and may you embark on that journey to deepen your business needs and not just for the sake of ‘AI’.

Editor’s Note: Tracy De Cicco is the CEO of Global Retail Technology Sales Consultancy, Konposit. She has 20 years’ experience in retail technology sales, including large enterprise software and professional services client sales and engagement. She has worked with many of the Fortune 10 global retailers and specializes in C-Level engagement, deal coaching, and resonant, effective sales messaging. She can be reached at tdecicco@konposit.com.