Where should retailers start with ‘generative’ AI?
Photo: Getty Images/simon2579

Where should retailers start with ‘generative’ AI?

“I think it’s exciting, what’s possible with generative AI,” Amazon.com’s CEO Andy Jassy said in a recent interview with Financial Times. “And it’s part of what you’re seeing with models like ChatGPT. But most large, deeply technical companies like ours, have been working on these very large, generative AI models themselves for a long time.”

Amazon is believed to be behind the “generative” AI race between Microsoft and Google, which hosted media events in early February highlighting their generative AI technology.

The AI-driven technology promises to understand and respond to natural language inputs with context. According to Investor’s Business Daily, Bank of America analyst Justin Post wrote in a note, “Use of AI will be a critical driver of all things Internet, including content relevance, ad performance, e-commerce conversion, marketplace efficiency and even customer service.”

ModernRetail said brands and retailers are looking to partner with ChatGPT-tech providers, with only a few larger, tech-driven corporations like Amazon, Shopify, JD.com and Alibaba seeking to develop similar tools themselves. Reuters wrote, “An exchange with AI known as a large language model likely costs ten times more than a standard keyword search.”

In the near term, experts see promising use cases for retailers that will incorporate generative AI into customer service chatbots, enabling faster, more relevant assistance.

Search is another likely focus, with many brands already significantly investing in AI-driven personalized product recommendations for years.

A study from Nostro found 69 percent of consumers head for the search bar first when visiting online retailers, but search falls well short of customer hopes in delivering personalized recommendations and relevant content.

“Generative AI holds significant promise for retailers,” wrote eMarketer in a report. “Not only can it power a better customer experience, but it could eventually become a more sophisticated form of predictive analytics, giving companies actionable insights into everything from the best locations for new stores to where to target investment. However, its applications in the short term will be limited as companies work out how to implement generative AI without risking customer ire or giving away too much information on the inner workings of their businesses.”

BrainTrust

"Generative AI is moving very fast and improving massively, but lets make sure it is fit for purpose before retailers risk their brand on pushing too fast. "

Andrew Blatherwick

Chairman Emeritus, Relex Solutions


"In the short term, retailers should focus on generative AI’s current, known and validated core strengths: search, data processing, and ideation."

John Lietsch

Chief Operating Officer, Bloo Kanoo


"We’re still a long way from AI being able to 'replace the humans'⸺and really that shouldn’t be the goal."

Nicola Kinsella

SVP Global Marketing, Fluent Commerce


Discussion Questions

DISCUSSION QUESTIONS: What advice would you have for retailers exploring generative AI systems like ChatGPT? Where do you see as possible near-term use cases?

Poll

15 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Dion Kenney
1 year ago

There will be many great suggestions for online chat, writing emails and advertising content, or developing product copy. A different application could be for internal use – provide salespeople with simulated sales training and discussion points for providing a better customer experience.

Andrew Blatherwick
Member
1 year ago

Generative AI is moving very fast and improving massively, but lets make sure it is fit for purpose before retailers risk their brand on pushing too fast. It is certainly going to be a big help in customers interfaces where simple tasks are required, however customers will get irritated very quickly if it does not deliver the results and expectations they have. As to new store site finding and investments, we have some way to go!

John Lietsch
Active Member
1 year ago

In the short term, retailers should focus on generative AI’s current, known and validated core strengths: search, data processing, and ideation. As generative AI propagates through online search, retailers need to understand and adapt to ensure their content remains relevant. Companies should also be able to apply AI’s increased data processing capabilities to improve existing data processing systems and, as costs for the technology normalize, retailers should be able to train generative AI tools on their own data to ensure better results and customer experiences. Finally, there’s no denying that generative AI can immediately help with ideation at all levels of an organization from writing ads to job and product descriptions. However generative AI is still in its infancy and is not at a stage in its evolution where it can be trusted to mind the store (yet).

David Spear
Active Member
1 year ago

I’d recommend leveraging AI tools in supply chain and customer service, and then expand into other areas of the business as these tools mature and become more ubiquitous.

Mark Self
Noble Member
1 year ago

My advice is to wait. As exciting as this technology is, I believe it is not yet ready for mass commercialization. Assuming the retailer has the resources I would assign a small team to monitor the technology in order to consider how to appropriately use it when it is “ready.”

Peter Charness
Trusted Member
1 year ago

The internet is massively funded through paid search and targeted advertising. So an unbiased AI-based search to get a consumer to a website is going to have to support some kind of revenue model. Once a consumer is on a retailer’s website, I have to wonder if there is enough data (or internal skills) to train a generative AI model from within the retailer’s data. So for most retailers I think there’s a “watch and learn” phase in progress here.

Nicola Kinsella
Active Member
1 year ago

ChatGPT is great for getting over writers block and getting ideas for ad, product, or other marketing copy. But it’s important to remember that it’s generated from all the other content on the web. Based on my own experiments evaluating it for my own team, I’d recommend the following:

  1. Create guidelines for acceptable use of ChatGPT generated content for your organization. You don’t want your teams to just copy and paste from ChatGPT directly.
  2. Edit for readability, so you don’t make your readers work too hard to consume your content, because so far in tests I’ve found it tends to write at a higher grade level by default.
  3. Edit for differentiation. If you don’t add your own “secret sauce” to content you’ll sound the same as everyone else and simply won’t be memorable.

Marketers have some great new AI-based tools available to them today. From image editing tools like Removal.AI that quickly removes the background from your subject (it even does a pretty decent job on curly hair) to content writers like ChatGPT. But it’s important to remember that successful marketing requires being memorable, creating an emotional connection, and having your own brand voice. We’re still a long way from AI being able to “replace the humans”⸺and really that shouldn’t be the goal. But if used the right way, these tools can increase your team’s productivity and elevate their work. That way they can spend less time on the mundane and more time on innovative creative and messaging that will disrupt markets and fuel your future growth.

Georganne Bender
Noble Member
1 year ago

Rich and I are emceeing the Independent Retailer & Sellers Conference this week where ChatGPT has been a hot topic, particularly among speakers who work with Amazon and other online sellers. Generative AI has already become part of the new content narrative.

Recently, I asked ChatGPT to write a bio for KIZER & BENDER and what came back was an interesting mix, some of which pertained to us but also to several other retail speakers.

This misinformation is what I have questions about, so I spoke with an industry attorney, asking who is liable if generative AI writes a blog post for a retailer and that blog post includes copyrighted material. The response was the retailer, so my advice for retailers exploring generative AI would be to use it responsibly.

Matthew Pavich
1 year ago

AI is definitely a worthwhile journey for retailers to invest in, but it is a journey and not a switch. As others have mentioned, most retailers would better benefit from investments in AI on the supply chain and pricing side as those are more proven with easily quantifiable results. When investing in AI capabilities that “communicate” in a more generative manner with humans, it is important to ask where and how it will be used. As a consumer, I personally don’t care if I chat with a real person or a robot — I do care, however, if they substitute light roast coffee for dark roast despite years of data showing my preferences. I also care if they are overpriced on an item that I can get somewhere else. In all cases, the right move for all retailers is to put their IT/data house in order to support the AI needs of tomorrow and to move along the AI journey regardless of where they are today.

Shep Hyken
Active Member
1 year ago

There is great opportunity with generative AI and ChatGPT technology. Near term: Some new companies are creating customer support solutions using ChaptGPT tech. It’s simple. Take a look at chat conversations. What questions are being asked? What are the best answers to those questions? Put them into the knowledge base. Then let the ChatGPT (or similar tech) do its thing. When a customer goes to the website to find contact information, encourage them to try the new chat. The key to success is to have the best answers for your customers’ questions. And be sure to create a seamless transition to live chat if the customer’s comments indicate they aren’t getting the information they need.

Questions about sizing, quality, reviews, nutrition, and much, much more are going to be incorporated into the new technology. Really, the sky is the limit.

Ricardo Belmar
Active Member
1 year ago

We’re still in the early stages of developing use cases for generative AI and I strongly believe we are more limited by our current way of thinking about retail challenges and how to apply generative AI than we are limited by the capabilities of the AI itself.

I see three strong areas for retailers to evaluate and invest in:

1) Customer service. Let’s face it, when surveyed, most consumers dislike engaging with chatbots. generative AI can bring a marked improvement to this area in leaps and bounds that can make those “old” chatbots obsolete. I expect we’ll see much experimentation here by retailers working with their tech partners.

2) Product discovery/search. How consumers search for a product is about to completely transform in the digital space. Today’s product search is based on consumers second guessing what keywords a retailer used to label their product catalog online. With generative AI, consumers can just ask, conversationally, any way they like to describe what they are looking for and include as many variables and variations as they want in that description. That’s very different from how consumers are searching on Amazon and Google today. Just think about how recommendation engines can improve in this context!

3) The store associate experience. Frontline store team will be able to help customers in new ways. Instead of hunting through online product info to answer a customer’s question, imagine a store associate speaking into an IoT device and conversationally asking the customer’s question to the AI. Then, take the response and continue the conversation as if speaking to a product expert. This scenario is not far away at all!

I see all three of these areas as near term and I am sure there are many more possible. Again, I would argue we are all so focused on how we have defined today’s retail challenges, that we try to put generative AI solutions a box to try and solve those challenges. We need more time to embrace and experiment with this technology to really define those outside the box solutions that can be transformative to the shopping experience.

thehonrosie
1 year ago

Consumer expectations for chat-based solutions will rapidly become as normal as shopping on your mobile. Retailers need to take note of ChatGPT and similar technologies; just assuming their customers “don’t like chatbots” based on experience of the basic options widely available today is to vastly underestimate the capabilities of generative and other forms of AI.

Oliver Guy
Member
1 year ago

Disclosure — I work for Microsoft.

This question is coming up a lot — everywhere. What is key is that output of generative AI is reviewed by humans — this avoids any issues. Guard rails and guidelines must be put in place.

There are multiple places where it can be experimented with — idea generation, marketing copy creation, creation of product descriptions (Carmax in the U.S. are using this for generation of car reviews) — I have even seen examples of changing product descriptions on websites based on segments or individual customers.

Christina Cooley
1 year ago

Retailers need to be exploring how to use AI now to run their businesses more effectively and to better serve their customers. First, it is important to understand the capabilities and limitations of the technology. These tools can definitely support customer service organizations with generating natural language responses. However, to get to that point, retailers need to invest in the necessary time to run “training” data through the system, conduct regular testing, and continue to evolve iterations to maintain the value that AI brings to the retail environment. Ultimately, incorporating an AI system is a significant investment, and retailers need to feel confident on the ROI and benefit to the customer in implementing.

Georges Mirza
Member
1 year ago

I think one of the early use cases of generative AI can be around data decisioning as part of the assortment review process. This is an area I am tackling in my Innovating Category Management series, first is assortment, and soon in New Age of Retail Insights where ChatGPT can be used to textualize trends in easy-to-read and understand paragraphs.