How will AI transform the online experience?
ShopBot Beta – Source: eBay

How will AI transform the online experience?

According to a survey from SLI Systems, 54 percent of mid-size retailers are using or plan to use artificial intelligence (AI) as an online tool in the next 12 months. The most popular applications are expected to be personalized product recommendations, customer service requests and chatbots.

The online survey of 234 e-commerce professionals primarily in the U.S., Europe, Australia and New Zealand showed that 56 percent are either using or planning to use AI for personalized product recommendations. That was followed by customer service requests, 41 percent; chatbots, 35 percent; and visual search, 32 percent. VR/AR, voice-activated apps and virtual buying assistants scored lower.EPIC Report Q3 2017 AI PlanningIn a note from late September, according to Barron’s, R.W. Baird’s Colin Sebastian indicated that the “overriding theme” at the Shop.org conference in Los Angeles was “the enormous change still coming to e-commerce/retail, driven by: AI/machine learning; new business models in the age of Amazon; shift to voice/conversational commerce and adoption of AR/VR.”

Noting that AI chatter had risen significantly at e-commerce conferences over the last two or three years, Mr. Sebastion said the message being relayed was that “machines will learn about and communicate with individual customers.”

At Shoptalk Europe in Copenhagen earlier this month, RJ Pittman, eBay’s chief product officer, gave a review of ShopBot, eBay’s personalized shopping tool “powered by artificial intelligence (AI), cloud computing, and human judgment.”  Launched last October on Facebook Messenger and later incorporated with Google Home, ShopBot enables consumers to text, talk or take a picture to search for items. The bot, which also uses historical purchasing data, then asks questions to fine-tune the search. 

Offering an example of a search for a dress for a wedding, Mr. Pittman, according to Essential Retail, said at the conference, “It needs to know the weather in Italy and what weddings actually are if you want to answer that question in a way no search engine can do today.”

He believes bots will one day be designing products for consumers.

BrainTrust

"AI is hot and a great buzzword but not simple to do well. "

Jon Polin

Cofounder and President, StorePower


"...one certainty about the future of retail — consumers will be working with AI interfaces to make choices throughout the shopping cycle."

Julie Bernard

Chief Marketing Officer, Verve


"...personalization is the low-hanging fruit for retailers...I think we’ll see the more innovative retailers find more creative uses for AI than that."

Ricardo Belmar

Retail Transformation Thought Leader, Advisor, & Strategist


Discussion Questions

DISCUSSION QUESTIONS: Do you see artificial intelligence significantly transforming the online shopping experience in the years ahead? If so, in what ways? Which applications will take hold first?

Poll

28 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Kim Garretson
Kim Garretson
6 years ago

Relatedly, AI will also dramatically change the way retail marketers reach individual shoppers with individualized messaging, also improving customer experiences.

Here is IBM describing its partner, Equals3, and its IBM-powered Lucy AI engine: “Lucy automates labor-intensive tasks and converts data assets into a quickly-searchable source of insights—freeing marketers to focus on complex, higher-level functions. Lucy is a cognitive problem solver. Whether it be determining market segments, doing competitive or market analysis, media planning, writing a marketing plan, developing a marketing strategy, creating organic content—Lucy’s assistance helps with every task.”

Dr. Stephen Needel
Active Member
6 years ago

While it is certainly coming, AI (or, more likely, machine learning) is not going to transform the experience for years. The price tag and computing power required are slowing down adoption, although those are becoming less of a problem all the time. The question retailers should be asking is how much AI is needed and how much benefit is gained by increasing that amount.

Max Goldberg
6 years ago

AI can be a boon for retailers and consumers provided that the machine can correctly answer a consumer’s question the first time it is asked or the first time a consumer is looking for a specific product. Too much time is being wasted with websites delivering information that a consumer doesn’t want or need. I look to AI for skillful recommendations and accurate information.

Chris Petersen, PhD.
Member
6 years ago

E-commerce has conquered the last mile of delivery. Amazon is now leading the way with in-home and in-your-car-trunk options. So while there are tremendous strides being made on how to personalize the purchase delivery, personalization in the early phases of the shopping journey has been lacking.

Personalization is the current Holy Grail of engaging shoppers. Research indicates that customers will linger longer and purchase more when the experience is personalized to them and their lifestyle.

For a number of years, David’s Bridal has had an interactive personalization process for not only the wedding dress, but also planning the entire wedding. It is a good case study of engaging customers through personal interaction of choices and then making that seamless to all the stores.

The rapid advancements in personalization will come through AI on the web. AI has the ability to “learn” individual customer patterns, backgrounds and behavior … and then customize choices and offerings in ways the the customer may not even realize. The interesting question is, how much personal information will customers being willing to share as part of this process?

The best predictions are that the will be no single “killer app” and the best solutions will include the customer interactively at multiple touch points in their journey.

Dave Nixon
6 years ago

AI will start to creep into online retailing as the costs and the technology become more accessible. Right now it is still cutting-edge technology. The key for advancing the adoption is to ensure that your digital channels are connected and have the foundational ecosystem to leverage it. Personal recommendations driven by AI are useless without a very robust view of the customer’s behaviors and preferences (or you are making bad recommendations based on bad data), and the ability to publish the personal content (not knowing the best approach for getting the customer to convert is as bad as bad recommendations). Those two things involve good foundational digital technology. Retailers are still struggling with connecting disparate channels, let alone a cohesive interconnected infrastructure to provide that personal experience shoppers expect. It’ll take hold … it’ll just be a while before it does.

Ryan Mathews
Trusted Member
6 years ago

The simple answer to the first question is yes, followed by about one million exclamation points. Now, the answer to the second question is a bit trickier. I see two primary applications: more personalization by sellers and more buy-in from customers. To some degree or another these technologies have the potential to virtually — excuse the pun — cancel each other out. So to me, the real issue is access to data. If sellers continue to have access to unedited shopper data we will see personalization — or at least mass customization or micro segmentation — explode. If on the other hand shoppers opt to either retain their own data and use it to program buying agents and/or they “block” the use and/or sale of data by sellers, many manufacturers and more than a few retailers will find themselves functionally disintermediated from the whole selling cycle.

As to which features will develop fastest, the list in the article is pretty good, although I think given the current proliferation of devices, voice activation will be more common than virtual and maybe even augmented reality.

Dr. Stephen Needel
Active Member
Reply to  Ryan Mathews
6 years ago

Ryan – to expand upon a point you made regarding data. We need to remember that a retailer only has access to data that they are collecting when it comes to shopping. Amazon doesn’t know what I look at on Jet.com and vice versa. That makes personalization a bit trickier.

Ryan Mathews
Trusted Member
Reply to  Dr. Stephen Needel
6 years ago

Stephen — True, which is why I believe Amazon is trying to insert itself in so many facets of the purchase cycle. Also, can you imagine a consortium of data collectors? Who wouldn’t want to share with Amazon if it meant they could know what Amazon knows? Over the next few years we are going to learn what customers value more — convenience and personalization or privacy and choice. Should be interesting.

Neil Saunders
Famed Member
6 years ago

AI has a big role to play, initially in facilitating search and helping customers find products that they want or need.

The problem with traditional search is that it relies on the retailer and consumer using the same terms for products — an alignment that doesn’t always occur. Moreover, while there may be partial personalization, the results are not particularly intelligent or context driven.

AI cuts through both of these issues and makes it more likely the customer will locate the product that’s right for them. That means higher conversions and ultimately higher sales.

Beyond this AI has many more applications, but the search and find functionality seems to be one of the most immediate ways in which it can be used.

Jon Polin
6 years ago

AI is hot and a great buzzword but not simple to do well. If executed poorly, customers will hold the offending retailer responsible, not the software company providing the functionality. But it is coming … and fast. I see the earliest relevant uses being to simplify the path to purchase for consumers who have a good idea of what they want but need a shorter path to get to it. Only down the road will AI be effective at helping consumers discover new brands and products that the consumer may not even realize s/he wants.

Art Suriano
Member
6 years ago

I definitely see AI growing not just for online uses but everywhere. We are still in the early stages with AI, with many issues that need to be resolved and improved. Today there are times when customers become frustrated — if, for example, AI chat does not understand them, or other another service a customer is using is not functioning correctly. However, we have to give this exciting technology time to be fully developed. Think back to the dial-up early days of the internet and how frustrating that was compared to where we are today. Within 10 years, AI will be as natural to us in cars, in our homes, online, in stores and just about anyplace we can imagine. How we integrate with it and how that impacts our workforce is another topic. But one thing about technology is that you can’t stop it and it always continues to move forward.

Lee Peterson
Member
6 years ago

“Alexa, show me some skinny jeans.” What does Alexa show you? Whose brands? What styles? Etc. Think of that simple transaction. Given that, no longer do you have to market those jeans to only me. You’d better be marketing those jeans to Alexa (aka Amazon). To me, that’s the biggest change with AI, otherwise, it’s still an on-going consumer paradise, just at another level. Funny, we all used to say, “with the internet, you can shop in your underwear.” Now we don’t even do that. A robot will do it for us, anytime, anywhere.

Brandon Rael
Active Member
6 years ago

AI or as IBM refers to it “Augmented Intelligence” is slowly creeping not only into the consumer facing commerce systems, but also the front office retailer systems. The AI revolution is not a sprint, rather it’s a marathon, and should be treated accordingly while the technology, adoption rates, acceptance and corporate change management is in motion. We are experiencing a degree of personalization and seamlessness with several of the leading online apps; Amazon, Google and others. However, we just may not be completely aware that these technologies are already enhancing our experiences.

The changes will take years to come to fruition, and the best strategy for retailers is to dip their toes in, develop proof of concepts, test groups and then roll out these new AI-focused strategies once the company is culturally read to take them on.

With any innovation it will take human talent and experience to drive things. The AI revolution will be no exception.

Gib Bassett
6 years ago

The AI coming to market today can help midsize and smaller online retailers adopt methods long used by Amazon and other large retailers far along the analytics maturity curve. So in some ways it’s like achieving parity. I think it’s important to think about AI, and more broadly analytics, as powering a retailer’s relationship with its consumer throughout the journey — not just in the active shopping phase. A more holistic approach is today realistic given the number of options to accelerate use cases that fuel a satisfying journey across the pre-shop, shop and post-shopping phases. That wasn’t the case in the past where a broader view required a longer adoption time horizon that few people had time for.

Adrian Weidmann
Member
6 years ago

Artificial intelligence (AI) is quickly becoming the latest shiny object of conversation but it’s been evolving for years. A colleague of mine was on the small team that developed and implemented the first recommendation algorithms for Amazon. The power of AI is when it is integrated in the broader digital and brick-and-mortar shopping ecosystem. These Big Data solutions can be combined with “little data” solutions — solutions that are pragmatic yet change the status quo and existing business processes (e.g. RFID, inventory visibility and management — we always seem to find our way back to supply chain visibility!). Artificial intelligence and machine learning will define the invaluable platforms needed to succeed and meet the expectations of the the digitally-empowered shopper.

Anne Howe
Anne Howe
Member
6 years ago

Yes, AI is certainly marching forward and I see it as an advantage to those retailers who can execute well. IMHO, great execution must include human judgement. We are still and always will be a human, and therefore emotional, society.

Ed Dunn
Ed Dunn
Member
6 years ago

AI and chat bots existed before the web itself. Retailers should not assume AI is future technology, it is just that the hardware and technology is now available to implement in a more mainstream fashion.

One of the AI-related technologies I see missing is crypto-based which is the most important and relevant right now. AI will need a secured identity and the ability to create a unique digital signature that is secure to conduct transactions such as purchases and return facilitation and allow the customer to verify they are talking to an authentic AI and not a phishing AI.

In every movie about AI, the AI had to learn about the owner and the operating environment from home to a spaceship. Retailers cannot outsource AI or rely on a tech company — the AI has to learn everything about a particular retail operation while researching external patterns and practices.

One more point — AI are not silos. Some AIs can post to Facebook or Instagram as well as follow up on customers and literally pay for digital advertising space and self-monitoring the inventory and digitally sign a purchase order. I would not try to kick the concept of AI down the road — AI will impact retailers in sooner than five years.

Camille P. Schuster, PhD.
Member
6 years ago

Personalization is a consumer demand so the use of AI is likely to start there. To be successful someone needs to write good algorithms, test their assumptions, test the results and monitor consumer changes. Using AI is not as easy as writing a formula and assuming that it works now and will continue working for the foreseeable future. How many people are qualified to do this well?

Shep Hyken
Active Member
6 years ago

AI is already transforming the online shopping experience. We don’t need to wait for years ahead. The ability for the “machine” to personalize suggestions based on past purchases and searches is just one way. A good chat bot will give the consumer the feeling they are talking to a person. Take a look at what 1-800-Flowers is doing. They have GWIN (powered by IBM’s Watson), which is their AI powered “concierge” that engages with the consumer to find the perfect gift. The future is here, and it will continue to be refined and improved.

Julie Bernard
6 years ago

The results that SLI Systems has released square in many regards with recent findings in our own research — we found, this past July, that 33%–43% of marketing leaders are deploying or testing-to-deploy artificial intelligence, with bots leading the field and voice assistants coming in at the slightly lower end of the spectrum.

All of this points to one certainty about the future of retail — consumers will be working with AI interfaces to make choices throughout the shopping cycle. Retail leaders that want to succeed at capturing their customers’ spend throughout this transformation must apply the same lessons we’ve learned from most of our industry’s recent mobile advances — build valid experiences for the user and the device, emphasize context and relevance, and personalize for, predict for, and surprise your mobile audience with ideas and inspirations they would not have discovered on their own.

The good news is, AI is tightly aligned with the kinds of engagements that breed these experiences. It’s heartening to see adoption of AI interfaces on the rise in the retail space.

Ricardo Belmar
Active Member
6 years ago

Absolutely, AI will transform many aspects of shopping. One key area I don’t see discussed very often, and I expect there is a very skilled startup out there somewhere that will soon address this, is the AI/chatbot that represents the consumer. This isn’t an Alexa or Cortana voice assistant — this is an AI that works for you, the consumer, and engages with brands and retailers to find what you need and want without you doing the work. When “search” reaches this level of personalization, we’ll hear the sound of retail marketers everywhere throwing their hands up and crying “uncle” over how to win consumer mind-share! Or perhaps we’ll see Apple’s Siri become such an agent….

AI is also an interesting technology for retailers because on the one hand it can enable a smaller retailer to look and feel like their biggest competitor and offer unique value to their customers. But, at the same time, in the hands of those big retailers, AI could allow them to reach new heights of innovation by investing more money and resources into development than any small retailer could afford. Today there is a high barrier of entry in terms of cost and skill at developing AI solutions. Gartner has even predicted that in the coming years people with AI development skills will be the highest in-demand workers of all. Retailers will be competing for these people for their own AI programs.

In some ways, personalization is the low-hanging fruit for retailers to focus on — they know customers want this, and it’s become the poster-child for AI applications. I think we’ll see the more innovative retailers find more creative uses for AI than that.

Doug Garnett
Active Member
6 years ago

I remain cautious about AI. First, what is called AI right now isn’t AI, but use of algorithms with big data. Many vendors simply re-named their big data efforts to be AI to grab the trendiness of the term. (And I say this with some depth in AI knowledge stretching back into the 1980s when I was a senior software engineer in aerospace.)

Also, personalization works in direct mail and direct response email marketing work. No question. But in electronic retail? Even Amazon (the most sophisticated recommendation engine around) has only suggested products that fit me personally once or twice in 20 years … despite having 20 years of data on my purchase habits. Something clearly is missing.

My recommendation is that retailers dabble in these efforts, but not make them a focus of major spending. Despite “everyone” saying how great they are, quite often the biggest winners are the ones who see past acclamations of the masses to find what works better.

Martin Mehalchin
6 years ago

Pittman gave a similar talk last week at the Forrester Research conference in San Francisco. The majority of personalization in the market today is rules-based and not truly AI and, as Doug Garnett points out above, after 15 years of working on it even leaders like Amazon still can’t get product recommendations to work all that well.

AI based implementations like eBay’s ShopBot offer a new approach with the potential to deliver far greater consumer relevance and resonance than rules-based techniques have been able to deliver.

Ralph Jacobson
Member
6 years ago

AI-enabled chatbots are already transforming online retailing around the world from India to the U.S., today. AI is promoting the full line assortments of merchants, rather than human call center associates simply recommending the top sellers. There are myriad additional uses for AI in the digital shopping experience. Look around at the best sites currently, and you’ll see some great examples right now.

Cameron Conaway
6 years ago

This falls in line with the research J.P Gownder presented last week at Forrester’s CXSF 2017: 51% of companies, especially the marketing departments within, are either using or are planning to expand their use of AI.

While AI was indeed the “overriding theme” at CXSF, it was tempered by another statistic from Forrester: the customer experience declined between 2016 and 2017.

Ultimately, transforming the online experience means transforming the human experience, and many AI-powered solutions are falling short.

On this, Pittman’s words rang true at CXSF as well:

“Personalization, generally, sucks. I don’t understand why. I’ve been in this business for 20 years and I’m seeing little more than recommendations. There’s no longer an excuse.”

Cynthia Holcomb
Member
6 years ago

This is both the most simple and complex question I have ever commented upon as a contributor to RetailWire. Simply, recommending products to a human must encompass the individual sensory preferences of matching a product to the individual preferences of fit, look and feel of an individual person. Short of that, a linear search is a micro segmentation-grouping people into buckets. Ok I can accept that. What I cannot accept are technologies focused solely on segmenting people into groups, as in the past 20 years, versus preference matching people to products they love.

Seth Nagle
6 years ago

AI has the capability of enhancing the digital shopping experience, but I don’t believe the industry is ready for that. As you can see in the poll besides personalized shopping not many experts have plans to implement AI just yet in their strategy.

There are still many factors and variables when it comes to AI and I don’t think retailers are willing to take the risk of upsetting their customers due to an incorrect algorithm.

Regarding the future, as natural language processing is built out in the years to come it will have a significant impact on online ordering as shoppers will just have to say “Order me the best paper towels, I need them today” and if everything goes right the AI should be able to place the order for the correct item, delivered at the requested time without hassling the shopper with additional questions.

So in theory AI should be able to act as your personal shopper, that will select the optimal product based on your needs, and do it all with the convenience similar to just pushing an Amazon Go button.

Hilie Bloch
6 years ago

As more and more data is collected on shoppers, potential shoppers and the products themselves, more and more analytics can be conducted to determine when a buying funnel has actually begun or is about to, which marketing and merchandising tactics work and which don’t per buyer. We all know that one size does not fit all but the true way of adjusting the strategy relies on our ability to understand shoppers, their interests and their needs, and then target them when and where its most effective. AB testing becomes integral to decision making and is easily and seamlessly conducted for experimenting with new ideas and for finding the right targeting channels for the buyers.