Since ChatGPT took the world by storm this year, other forms of AI that have grown in popularity seem to have fallen into a bit of oblivion. But AI and automation capabilities will become even more essential in retail.
A 2017 McKinsey study found that only 20% of organizations reported using AI in their business, but by 2022 that amount will more than double. The World Economic Forum this year said, “AI services in the retail sector are projected to grow from $5 billion to over $31 billion by 2028.”
AI is becoming more and more necessary to remain competitive in the retail industry. AI in retail involves using automation, data, and technologies such as machine learning (ML) algorithms to deliver highly personalized shopping experiences to consumers. One important use of AI technology is in self-checkout innovation, where providing a safe scanning method can also help prevent shoplifting.
These AI retail platforms can run without human assistance, giving customers more control over the shopping process. As an incentive for loss prevention, a new system with AI authentication is used to log data on suspicious shoplifters.
Key ingredients for modern retail success
E-commerce merchandising teams are also busy trying to keep up with the demands of modern digital merchandising. Automating manual tasks with AI can save a lot of time.
Artificial intelligence can also be a competitive advantage. AI can provide data-backed insights to support more effective product recommendations, omnichannel engagement strategies, and long-term business goals.
But OpenAI’s ChatGPT code isn’t the only tool retailers should consider. ChatGPT doesn’t make other forms of AI unnecessary, according to Eli Finkelshteyn, co-founder and CEO of Constructor, a product search and discovery platform for retailers. Contrary to popular opinion, ChatGPT is not completely ruling out other forms of AI.
“If you ask ChatGPT itself, you’ll get the same response. We will create one,” Finkelshteyn told the E-Commerce Times.
Balancing AI tools for deep product discovery
There are certainly situations where talking to employees online can prove beneficial and appropriate. However, not every product you plan to buy will need it, he said. “Search is and always will be a better tool when shopping online if you know exactly what you’re looking for,” he argued.
If you just want to digitally browse the aisles and see the variety of shirts, electronics, and snacks a retailer carries, category pages are still a better tool. But some amount of the time (maybe he’s 5%-10%), you’ll want to talk to a specialized sales clerk and elaborate on what you’re looking for.
This is where ChatGPT can do a much better job in e-commerce than what is currently available, he acknowledged. That means we can close deals by understanding long-form queries, answering questions, and helping you find what you need.
Leveraging AI to deliver personalized shopping experiences
AI can rank and personalize search results based on what it knows about product catalog trends. You can also leverage what you know about shoppers from their behavior across your site and other channels of your brand.
So, for example, if someone searches for “shirt,” the shopper’s favorite brand, color, and price range may be prioritized. While doing this, it also optimizes the retailer’s KPIs such as revenue, conversions, and profits.
For example, Constructor’s platform uses AI and machine learning to gather details from every user query. According to Finkelshteyn, it can also include feedback given to the system hundreds of millions of times each day through post-query interactions with the results, allowing for continuous refinement of these results and experiences.
A conversation about the role of AI in e-commerce
The E-Commerce Times asked Eli Finkelshteyn to discuss what e-commerce companies can expect from AI in the years to come. He also knows the traps retailers should avoid when considering AI.
E-Commerce Times: How else can these other AI options continue to thrive in ChatGPT’s shadow?
Eli Finkelstein: ChatGPT is great, but it’s not the best tool for every AI problem. For example, it is not the best tool for domains that return large result sets and filters, such as general e-commerce search.
It’s also not the best tool for making good use of your site’s traffic and clickstream. [behavioral] Data for ranking all of these results, or learning appropriately from the e-commerce site’s specific user base and clickstream data to see how rankings change over time, or how specific shoppers Decide if it should be personalized for you. This makes sense, as these are not the problems ChatGPT is trying to solve.
Instead, these are uses where clickstream-based AI trained on the unique traffic of e-commerce sites will continue to be a better fit, allowing retailers to improve shopper behavior and create consistent omnichannel experiences. Case.
As advances continue on both sides of AI, what can retailers expect?
Finkelstein: The two promise an exciting future. Using ChatGPT and existing forms of AI, especially clickstream-based, opens up new possibilities for combining them to create amazing new user interfaces and experiences.
For example, if you need to explain on your e-commerce website that you are looking for a product for a specific purpose or event, such as clothes to wear to a friend’s engagement party or ingredients to make a casserole, that is currently difficult. The formats available for product discovery, such as searching and browsing category pages, are not great ways to ask or find out about this sort of thing.
But incorporating ChatGPT to help retailers understand shoppers’ questions in this long-form, abstract question brings innovation. Using clickstream-based product discovery AI to show trending, personalized results that match users’ needs is an attractive proposition.
How has ChatGPT’s popularity and technological advances impacted Constructor’s product search and discovery platform for retailers?
Finkelstein: Generally speaking, everyone is still trying to understand ChatGPT’s place in e-commerce. Finding the right use cases and interfaces requires experimentation and flexibility. In particular, it’s also important to make sure that his ChatGPT-based technology that he deploys to his users is really useful.
What is Constructor doing to drive innovation like this?
Finkelstein: We provide an AI-based platform for product search and discovery, helping retailers deliver highly personalized experiences across channels that also reflect KPIs. The impact of ChatGPT on our work at Constructor is that many of our retail partners are excited to try our prototypes in this space, and we will be integrating prototypes sooner than any previous new technology we have deployed. We are willing to invest development resources to
We have long believed that product discovery is much better than what shoppers are currently accustomed to. ChatGPT has made retailers more open to experimenting with new forms of product discovery and knowing what shoppers value most.
While conducting such experiments, how can retailers keep in mind the actual utility for shoppers?
Finkelstein: Unfortunately, many of the efforts we’ve seen so far in this space, such as the integration of ChatGPT with other retail technologies, have been mere gimmicks. It looks cool the first time you play with it, but it’s useless to shoppers.
One company has published a chatbot that uses ChatGPT, but basically it just lets users enter their needs in the form of a search to search for products. It’s actually less useful than just using the search bar. This doesn’t provide real value, so sites that implement this may find that shoppers may never come back after he’s tried once or he’s twice.
ChatGPT and the energy around it gives all of us in the product discovery space the opportunity to revolutionize the way we find products with new interfaces and technologies. Our goal at Constructor is to harness that energy and excitement to create technology that shoppers will want to use again and again, and not waste excitement on gimmicks that shoppers will only try once.
What can e-commerce companies expect from AI in the years to come?
Finkelstein: Expect lots of experiments. ChatGPT and, generally speaking, the technologies behind it (Transformers and Large Language Model (LLM)) enable many things that were not possible before. But just because something is possible doesn’t necessarily mean it’s worth it to e-commerce companies and their shoppers.
Experimentation is therefore expected to find the most worthwhile places to apply this new technology. Constructor already uses this under the hood to make the results returned to the user more contextual and attractive. But this is just scratching the surface.
What should retailers consider when deploying AI tools, and what traps should they avoid?
Finkelstein: Above all, make sure what you’re doing is likely to be worth it for you and your shoppers. There are already a ton of gimmicks built around ChatGPT that are disappearing, but there will be more to come.
So, think twice about the value ChatGPT and other AI-based technologies bring before committing. Also think about the problem or problem you are trying to solve.
This could be connecting data across channels, reducing manual work associated with merchandising tasks, or reducing search abandonment. Consider the best use cases for AI.
Where can ChatGPT add value, and where is clickstream-based AI a better fit? Where is the right place?
Valuable technology in this field arrives. But we also expect to see far more so-called innovations that look cool but aren’t as useful or worth the effort as moving the needle. Retailers who can tell the difference will be the most successful.
What do you see as the safety and ethical concerns associated with the increased use of AI? For example, several European countries recently banned the use of ChatGPT.
Finkelstein: Some forms of artificial intelligence, such as conversational AI, can confidently lie and deceive. Advances in generative AI, including those used in deepfakes, are making it increasingly difficult for people to trust their eyes and ears, and as technology advances, this problem will only get worse.
Should governments be more involved in overseeing AI technology?
Finkelstein: Governments cannot and should not rely on AI companies for legislation and must act to protect and educate their citizens as much as possible. As a company at the forefront of AI and in positions of power, it is important that we do the right thing and act ethically in our use of AI. But it is also very reasonable for the government to ensure that it does.
How might government intervention affect future AI developments?
Finkelstein: The flip side of this is that we live in a globalized world with internet access for billions of people in nearly every country. This makes it difficult for AI to crack down. If governments delay AI research by companies under their jurisdiction, it may help protect citizens in the short term, but it could hurt them in the long term.
Rapid advances in AI will give companies in countries that fail to make the same decisions a head start and a significant advantage. We are still connected to the internet. AI developed in less careful places will still affect people around the world.
Finding the right balance will be a major challenge for world leaders, and I hope our leaders will tackle it. In the meantime, we, the leading AI companies, should always consider what we are doing, make sure it is ethical, and encourage our colleagues to do the same. is recommended.
Editor’s Note: For more information on issues related to regulating or pausing AI development, see The AI revolution is at a tipping point.
