Operations
13 January
Google debuts 4 AI tools for retailers
Shelf checking, online window shopping and product recommendations are the focus areas.
Photo by Alex Dudar on Unsplash
×
The Current, delivered daily.
Ecommerce news, right to your inbox.
Sign up for our email newsletter.
Subscribe.sticky-cta-wrapper {
width: 100%;
position: fixed;
bottom: 0;
left: 0;
background-color: var(--black);
}
Subscribe to The Current newsletter
The NRF Big Show arrives in New York on Jan. 14, and that means companies will be rolling out their latest technologies and partnerships to start off 2023 strong.
Google is getting a head start. Just ahead of the event, the company’s cloud division shared details on four new AI-powered releases for retailers on Friday. The technologies are designed to advance in-store shelf checking, as well as ecommerce browsing and recommendations.
The launches show how Google Cloud is expanding a suite of products for others. This is distinct from Google Shopping through the search giant’s own site, which has also gotten plenty of upgrades over the last year.
"Upheavals over the last few years have reshaped the retail landscape and the tools retailers need to be more efficient, more compelling to their customers, and less exposed to future shocks," said Carrie Tharp, VP of retail and consumer at Google Cloud, in a statement. "Despite uncertainty, the retail industry has enormous opportunity. The leaders of tomorrow will be those who address today's most pressing in-store and online challenges with the newest technology tools, such as artificial intelligence and machine learning."
For those looking to see how AI will play a role in commerce, Google’s technologies offer a few use cases, from store operations to personalized online shopping.
The technologies include:
A shelf-checking solution
This aims to ensure that shelves in stores don’t become low or empty. Built on Google’s Vertex AI Vision, this uses Google’s database of people, places and things to help recognize products. It uses two AI models: a product recognizer and tag recognizer.
The goal is to solve a particular problem: “how to identify products of all types, at scale, based solely on the visual and text features of a product, and then translate that data into actionable insights.”
The AI tool can identify products based on images taken from a variety of different angles. Currently in preview, the product allows retailers to retain their imagery and data.
An online window shopping tool
This feature is designed for browsing and product discovery.
Here’s how it works: Shoppers choose a category, such as “women’s jackets.” Then machine learning selects optimal ordering of products in the results. The system, which is powered by Discovery AI, learns the ideal order on each page, and improves over time with more data.
The feature can be used on browsing, brand and landing pages.
Typically, products are ordered based on best-selling items, or human-written rules. The AI tool essentially self-curates a selection. For retailers, the potential benefits are more revenue per visit and team time saved.
Personalized search and browsing results
This is designed to show customers products that align with their tastes and interests.
AI is employed to analyze a customer’s behavior on an ecommerce site, such as their clicks, cart, purchases and more. This helps to determine shopper tastes and interests, and AI then moves up products that match them in results.
Google said the results are based solely on an interaction on a specific retailer’s site through an account or first-party cookie, and not linked with Google more widely.
Upgrades to Recommendations AI
This is designed to help retailers arrange and coordinate recommendation panels that appear on ecommerce sites. New capabilities include:
- Page-level optimization to help an ecommerce site dynamically decide what product recommendation panels to show shoppers.
- Revenue optimization uses machine learning that helps to identify recommendations that will increase revenue per user session.
- A third tool, built with DeepMind, uses machine learning to analyze product categories, item prices, and customer clicks and conversions. It finds recommendations that balance customer satisfaction and revenue lift.
- Buy-it-again, which provides recommendations for repeat purchases based on a shopper’s history.
Subscribe to The Current Newsletter
SubscribeTrending in Operations
Shopper Experience
08 June
AI, but make it fashion: Rent the Runway, Stitch Fix look generative
"Fashion ecommerce is one of the most cumbersome customer experiences that exists," said Rent the Runway CEO Jennifer Hyman.
Photo by Daniel G. Wells III, used under a Creative Commons license.
The rise of generative AI is bringing with it a groundswell of interest and concern about how the capability to automatically synthesize information and create something new will change how we work.
Given that AI will sit within the architecture of our digital lives, it’s also worth considering how the technology will introduce new tools for other aspects of life, as well.
For two ecommerce innovators in the apparel space, it’s a time to explore how it will transform shopping. Rent the Runway is set to roll out new AI-powered search capabilities, while Stitch Fix is drawing on a long history with data science and machine learning to personalize the inventory buying process.
Here’s a look at the initiatives underway at each company, and their visions for the future:
Rent the Runway: From search to concierge
Rent the Runway is putting a focus on the customer experience this year as it seeks to retain more subscribers and continue a yearslong push toward profitability.
This is resulting in the introduction of a variety of new initiatives, from the addition of an extra item to all orders to speeding up page load times. Yet as CEO Jennifer Hyman zooms out, she sees change being necessary on an industry-wide level in fashion. Beyond adding new features, AI can play a transformational role.
“I think that fashion ecommerce is one of the most cumbersome customer experiences that exists. You are searching through pages and pages and pages of content to find the items that you like and no one likes doing this,” Hyman told analysts on the company’s earnings call this week. “As an industry that still is selling physical products, AI is going to be -- fashion is going to be a major beneficiary as an industry.”
As a rental service, Rent the Runway has a distinct niche in fashion that lends itself to AI’s advantages, Hyman said. As opposed to a retailer that a consumer may visit a couple of times a year, RTR is used frequently by customers. So Hyman said there are opportunities to turn Rent the Runway into a “utility” by creating a more seamless experience.
This frequent use also provides a “highly unique” dataset, Hyman said. They know what a customer is planning to do based on what they rented. They know whether she liked or disliked an item, and many customers are reviewing 10 items per month. They know her size and how an item fits. This can be put to work in tools that allow customers to ask questions, and find answers.
The first application that combines AI and these advantages will appear in the coming weeks, when Rent the Runway plans to launch a beta of AI-driven search. The tool will allow customers to search for common terms or use cases for an item. So a person will be able to write “Miami vibe,” “‘clambake in Nantucket,” or “tropical motifs,” and receive results about what to wear for such an occasion.
The goal is to help customers sift through the endless aisle, and instantly finds what's right for them.
“I think that across all fashion sites, all over the world, the way that people are searching for product is fairly vanilla, it's fairly functional, right?" Hyman said. "You can go to a site and search for a T-shirt, you can go to a site and search for a black-tie gown. The fact that we're going to be able to enable our customers to search how they actually want to use this closet in the cloud, to search for items to wear to my beach bonfire this weekend, that is a completely different way to search, and I think that it really brings out the value proposition of what a closet in the cloud is all about."
Hyman sees this as a first step in the company using AI models to improve the product experience, and expects more tools to appear in the coming months. RTR is also introducing an SMS concierge experience for onboarding that allows customers to text with a member of the customer service team. The company is already exploring ways that AI can be incorporated into that tool, as well.
In the longer term, Hyman said the company has a vision that will leverage AI to allow customers to communicate with Rent the Runway asynchronously across different modalities, and have a stylist that is constantly available to recommend items, pick out new inventory and answer questions.
“If we are utilizing AI appropriately over the next few years, I see no reason why someone even has to come to our website,” Hyman said.
Stitch Fix: Inventory buying and beyond
Stitch Fix has long married AI with human curation to provide outfits on a subscription basis.
“For years, we have utilized capabilities in generative AI, injecting scores and language into our personalization engines and, more recently, automatically generated product descriptions,” CEO Katrina Lake told analysts. “We have also developed and implemented more advanced proprietary tools such as outfit generation and personalized style recommendations that create a unique and exciting experience we believe is unmatched in the market.”
A new area where the company is applying AI is inventory buying.
“We have historically utilized a number of tools to make data-informed decisions with our inventory purchases,” Lake said. “Now, directly leveraging our personalization algorithms, we have developed a new tool that creates an exciting paradigm shift, which will utilize math scores at the client level to drive company-level buying actions. We expect the clarity of demand signals at the individual client level to drive more proactive and efficient inventory decisions as a company. And because of this, we expect to see higher success rates on fixes and drive increases in keep rates and [average order value] over time.”
Early results are promising. When compared with existing buying tools, testing showed a 10% lift in keep rate and AOV. By the end of this quarter, Stitch Fix expects 20% of all purchase orders to be algorithmically informed.
With experience using AI and a team in place to build, Stitch Fix is investing in the technology. Like Rent the Runway, it also has a unique dataset that offers an immediate advantage.
Here are Lake’s thoughts about how Stitch Fix’s AI strategy:
One of the things that I love about our experience is that we have generative AI that's really in more of a visual format. And so, the outfits that we have in our app, those are actually taking into account your preferences, what we know about you, and then in combination with what we know that you own in your closet. And to be able to kind of continue to push that technology and to be able to continue to give people more value in their experience with Stitch Fix, that's a really good example of, I think, a capability that is, firstly, really aligned with our capabilities around data and personalization and really unique to us.
And then I think it's also really compelling because I really think that pushes us as we think about what that addressable market is. I think if we can push outfits to be something that can be an asset to everybody, I think that is a universal thing that people would love to be able to have, is to have access to advice on a daily basis around what to wear and how to wear it.
While these are distinct companies, their plans lead us to a common conclusion: While the talk around generative AI might be new, many technology-forward companies already have assets sitting inside them that can be leveraged to build new tools. Uncover what’s already there, learn about the AI’s capabilities and develop a solution that's right for your organization. Then, talk to customers to determine how to improve it. It might mean commerce looks different, but that’s okay. The point is to create a better experience.
Keep reading...Show less
Loading...
Subscribe to The Current Newsletter
SubscribeLatest from Operations
.sticky-cta-wrapper {
width: 100%;
position: fixed;
bottom: 0;
left: 0;
background-color: var(--black);
}
Subscribe to The Current newsletter