Amazon faces a dilemma — fight AI shopping agents, or join them
Robinson | Getty Images
Amazon CEO Andy Jassy could see how dramatically AI was changing e-commerce.
He told employees in June that AI agents would begin to infiltrate various aspects of daily life, “from shopping to travel to daily work and tasks.”
Four months later, Jassy said in an earnings call that Amazon expected to partner with third-party agencies and was in talks with some providers, though he did not name them.
Amazon is now looking to hire a leader in corporate development who will help build strategic partnerships in areas including “agency business.” a new job posting.
The rapid evolution in Amazon’s perspective on AI-powered commerce underscores how quickly online retail is changing and the risks the company faces if it doesn’t act aggressively to maintain control over its future.
The company followed as OpenAI, Googlesurprise and Microsoft In recent months we have released a number of e-commerce agencies aimed at changing the way people shop. Instead of visiting Amazon, Walmart or Nike Consumers can rely directly on AI agents to do the hard work, like scanning the web to find the best deal or perfect product and then purchasing the product without leaving the chatbot window.
The first shopping agents from AI leaders were launched about a year ago. consulting firm McKinsey He predicted that agency commerce could generate $1 trillion in U.S. retail revenue by 2030.
It’s a trend that poses a threat to Amazon’s margins and relationships with its customers. When a consumer uses ChatGPT to initiate a purchase (for example, OpenAI) collects a “small fee” from each transaction.
“With an agent on ChatGPT, retailers run the risk of foregoing transactions on their own site in order to pay fees on someone else’s toll for the same transaction,” Sucharita Kodali, a retail analyst at Forrester, said in an interview.
Some companies are trying to find a middle ground between working with agency providers and competing against them. Walmart, Shopify and others have adopted a strategy of frenzy, announcing partnerships with AI companies while continuing to develop their own tools and building guardrails around how brokers can access their sites.
Shopify CEO Tobi Lutke wrote: Publish on X He said Tuesday that his company is “building all layers of infrastructure to power a new Cambrian explosion of creativity in shopping.”
“I am really excited about Agency Trading,” Lutke wrote. “There are so many great things being built. Everything I test feels so nice and right.”
Amazon is playing defense so far.
The company recently updated the code underlying its website to prevent external AI agents from crawling the site, as part of an effort to separate valuable training data from competitors. As of Tuesday According to its website, Amazon blocked 47 bots, including those from all major AI companies.
Amazon even took the issue to court. In November, Amazon sued Perplexity over an intermediary that allowed the startup to make purchases on behalf of a user in the Comet browser. The company claimed that Perplexity took steps to “disguise” its representatives so they could continue scraping Amazon’s website without its approval.
Perplexity called the lawsuit a “bully tactic.”
Meanwhile, Amazon is investing heavily in its own AI products. The company released a shopping chatbot called Rufus last February and is currently testing an agent called Buy For MeAble to purchase products directly from other sites in Amazon’s e-commerce application.
Personalized shoppers
Morgan Stanley predicts that by 2030, nearly half of American consumers will use AI agents, and this technology could contribute $115 billion to US e-commerce spending.
“We believe agency trading – essentially the ability to have a personal digital interactive shopper – will be the next most important GenAI-powered unlocking method,” Morgan Stanley analysts wrote in a report in November.
They noted that a mid-single-digit percentage of consumers currently begin their “purchase journey” through AI, but that could increase over time, as roughly 40% to 50% of Americans currently use AI for product research.
Traffic from AI chatbots to US retail sites has increased in recent months, especially during the holiday season, but research suggests Google search still outperforms in terms of conversion rate and revenue per session.
Artificial intelligence-supported shopping remains a nascent market.
OpenAI’s Instant Payment tool, which launched on ChatGPT in September, is only available on some products sold by Walmart. Shopify, Aim And Etsy. Users can only purchase one item at a time and cannot connect to loyalty memberships like Walmart+.
Agents are also prone to glitches.
Scot Wingo, founder of e-commerce software startup ReFiBuy, recently tested Perplexity’s Instant Buy tool, which allows users to purchase products directly from the search engine.
Wingo tried to buy a cable-knit sweater Abercrombie and Fitchbut Perplexity’s representative kept giving error messages even though both items were in stock on the retailer’s website. Finally he gave up.
Earlier this month, Wingo suggested a Breville espresso machine while searching for a coffee machine on ChatGPT. When he clicked on the product, he was surprised to see the image of a garden rake.
“These scanners go out, they get this data, and you never know exactly what they’re going to get,” Wingo said.
‘The leader’s dilemma’
While Amazon considers its next move with shopping agents, it has been quietly allowing them access to some of its properties.
The robots.txt files of subsidiaries such as shoe retailer Zappos, fashion site Shopbop, and deal site Woot do not contain any language blocking agents that determine how browsers can access certain web pages.
“A lot of times they will use subsidiaries to experiment,” Wingo said. “Zappos has its own experience and database, so they don’t let all the horses out of the stable.”
If the company decides to allow agents access to its primary e-commerce platform, it could ultimately take a page from its competitors. Shopify And Walmart We have set limits on what external shopping agencies can and cannot do on their sites.
Wingo said Amazon might be willing to let agents access its catalog, but it likely wants to protect more valuable data than its rivals; such as extensive customer reviews and sales rankings, both of which indicate the quality of a product and can help improve an AI chatbot’s responses.
“Those are probably the two most private data points that I would want to protect if I were Amazon,” Wingo said.
Amazon does not give up on the tools it develops.
Rufus’ capabilities have improved since Amazon first launched it last year, and the company is rolling out the chatbot to more areas of its site to encourage user adoption.
Amazon recently added a feature Where Rufus can automatically purchase products on behalf of the Prime customer when a certain price is reached. The chatbot now recommends products from sites across the web, not just Amazon.
Amazon also began testing a feature in recent weeks that allows Rufus to create exclusive shopping guidesIt’s similar to OpenAI’s “shopping research” tool, which was released last month.
“Instead of the innovator’s dilemma, I would say Amazon is in what I would call the leader’s dilemma,” said Jordan Berke, founder and CEO of retail consulting firm Tomorrow. “Their market share is so important that they have the most to lose.”
WRISTWATCH: How did Amazon dominate the US clothing market?
