Can Amazon AI voice replace customer reviews? It’s starting to try

Anyone who has ever gone in the hunt for a product evaluate on Amazon is aware of how priceless the expertise of different customers might be, and the way straightforward it’s to fall down the rabbit gap of customer feedback, from five-star raves to one-star takedowns — typically tons of of phrases to get to the purpose. As Amazon continues to roll out AI options it says will make purchasing simpler, AI-generated audio descriptions of merchandise are within the combine and scraping that customer commentary — possibly in the end settling right into a place to replace it as a go-to supply of shopping for info. 

Called “Hear the Highlights,” the AI-voiced product descriptions depend on a big language mannequin to script the abstract primarily based on a wide range of sources, pulling from Amazon’s product catalog, customer evaluations, and data from throughout the net, after which translating the content material into short-form audio clips. The summaries started rolling out throughout the summer time on choose merchandise to a subset of U.S. clients and have now reached all U.S. clients as a button within the cellular purchasing app, scaling up to cowl over a million merchandise.

A singular enchantment of Amazon, and different e-commerce choices, has all the time been the flexibility to get info from precise customers, not simply product descriptions. Of course, bogus reviews have long been a problem for Amazon, regardless that it bans use of paid promotion and different inauthentic types of evaluate writing. You nonetheless have to type by means of varied efforts to recreation the system, and in recent times, indicators that the latest evaluate writers are chatbots like ChatGPT. But precise individuals sharing their precise, idiosyncratic expertise with a product — clients as a supply of data, and knowledgeable resolution making — have been a key a part of the training course of from clothes and shoe sizes to security and extra slender questions: When buying a brand new toaster, does it do bagels, is the timer setting correct, how straightforward is it to clear the crumb tray?

Can AI enhance on that? Reviews are by nature unwieldy within the mass individuality, however human readers have confirmed fairly adept at distilling what they want from the human chaos. In some sense, the mind of the common Amazon evaluate reader is a reasonably good giant language mannequin, skimming and choosing out key phrases and key info, so an AI may have to do it at the least in addition to a human would if it is to add to the customer expertise as promised. 

AI and cognitive overload

AI does have its benefits. For one, it will not expertise cognitive overload when going through an electrical tea kettle on Amazon with over 32,000 evaluations. It can comb by means of the info, however can it offers us solely what we want or need to know? That can nonetheless be difficult for AI to get proper, and it could be inherently problematic to combine product catalogs, customer evaluations, and internet info right into a single distillation – with the sources of this info every coming with its personal set of intentions.

“It’s important to recognize where AI is currently strong, such as in automation and pattern recognition, and where it still falls short, like in judgment-heavy tasks,” mentioned Ankur Edkie, CEO & co-founder, Murf AI, which creates AI voiceovers. “A key question is whether there’s a way to factor in customer context as an input while generating these summaries,” he mentioned.

The worth of AI, in accordance to Edkie, is discovering the right problem-capability match. If that is not achieved, the sense of gimmickry is probably going to sneak in by means of a door left open for AI fatigue, which he says shoppers are seemingly experiencing by now.

Amazon’s Hear the Highlights AI-audio summaries are at the moment the identical for each person.

“The AI summary needs to capture nuance and context. For example, even a few negative reviews on safety can outweigh many positives on other features,” Edkie mentioned, including that if a customer is concentrated solely on product efficiency, then summaries that emphasize worth may not be related. 

The capability to consider context, and to make the evaluate course of extra of a dialogue between client and machine is probably going the place the expertise is headed – in different phrases, towards the agentic side of AI, the place Amazon can be actively including to its AI commerce instruments, corresponding to Rufus, and a purchasing device known as Interests AI, which prompts customers to describe an curiosity “using your own words,” after which it generates a curated collection of merchandise. That characteristic, rolled out within the spring, is separate from the principle search bar on Amazon’s web site. 

Chat and audio summaries will stay among the many methods to interact, however having real-time conversations with an AI voice agent — asking particular questions, clarifying considerations, and getting deeper insights from evaluations — is what’s going to shift the expertise from one-way supply to two-way discovery, making it way more customized, in accordance to Edkie. “Currently, you can interact with Rufus through text or by using voice input, but Rufus cannot talk back, its responses are text-only,” he mentioned. “With voice agents, however, you can have a two-way conversation with a bot that speaks to you,” he added. 

For now, one phase of consumers seemingly to see quick advantages is visually impaired shoppers, making accessibility an intriguing facet of the characteristic, however the voice have to be high-quality and ship the content material precisely.

Brian Numainville, principal at client analysis agency Feedback Group, says by offering an audio-based different to visually introduced info, these kinds of options have the potential to make purchasing extra accessible, changing detailed textual content into simply consumed audio summaries. However, for it to really profit individuals with visible impairments, the characteristic have to be thoughtfully designed. According to Numainville, this would come with making certain full compatibility with display screen readers and keyboard navigation, offering clear, structured, and concise summaries, and avoiding overly lengthy or complicated audio displays. The high quality and readability of the AI-generated voice may even considerably impression usability. 

“The shift from diverse human reviews to AI-generated summaries might mean losing important nuances, context, and personal touches,” Numainville mentioned.

Risk of dropping distinctive shopper insights

The tendency of AI to concentrate on frequent themes can dilute responses even because it distills them.

Human evaluations have a tendency to embody digressive stories and particulars round extremely particular use circumstances — consider what motivates somebody to write a evaluate within the first place, and the way this pairs effectively with the patron’s anxieties and decision-making course of — all of which will help to make the purchasing expertise extra private and insightful. 

“AI might overlook unique insights or niche needs that don’t align with the majority of responses,” Numainville mentioned. “Additionally, the ability to critically interpret reviews — like spotting biases or trusting certain reviewers — is diminished with AI summaries.”  

Scraping quite a lot of content material to generate the summaries additionally sacrifices the annotation that may distinguish amongst product descriptions, customer evaluations, and internet info.

“It’s not 100% clear how much of the product description versus reviews is used in the current implementation — Amazon hasn’t fully detailed this mix,” mentioned Numainville. 

Amazon declined to remark, referring CNBC to publicly out there info on the characteristic.

“Product descriptions are typically marketing-driven and emphasize positives, whereas reviews reflect real user experiences and include both pros and cons,” Numainville mentioned. Mixed with out attribution — which by definition flattens info sources — it may very well be tough for patrons to discern ad-speak from real evaluations, with the outcomes be one thing like native promoting.

“This blending, if it occurred, could unintentionally mislead consumers by lending factual authority to subjective opinions or disguising promotional material as unbiased reviews,” Numainville mentioned. 

Research does present that clients are inclined to trust voice as a format of data supply, no matter how balanced the data being delivered is.

“It seems likely to me Amazon’s desire to sell products would weigh more highly than incorporating all critical perspectives on a product,” mentioned Tama Leaver, professor of web research at Curtin University in Australia.

Another concern is that “AI might weigh average and overall scores, while buyers often look at the few negative reviews – the one stars – even if there isn’t a lot of them,” Leaver mentioned.

Dr. Nauman Dawalatabad, a analysis scientist at Zoom Communications, mentioned in his private view the expertise is transferring within the route of higher customer expertise. “I take it as technology helping us to make informed decisions,” he mentioned, citing the cognitive fatigue and wasted time that may come from studying by means of customer evaluations.

If a streamlined description finally ends up main to impulse purchases and regrets, that’s on the client, he says, and no totally different than the best way issues have all the time operated — the identical expenses of delicate coercion may very well be made towards all advertising and marketing efforts. He thinks that as voice-based agentic AI continues to evolve and shoppers begin speaking (as a substitute of typing and looking out) with an AI agent and describe what they need, “it will get you exactly what you need.”