GenAI
©Tony Studio via Canva.com

June 5, 2025

Will GenAI ‘Become the Consumer,’ Acting as a Second Self When Buying on One’s Behalf?

Generative AI, or GenAI, is a nonstop topic of conversation in the retail business, whether the conversation dwells upon the recent rise of agentic artificial intelligence or the increasing abilities of AI to streamline supply chain operations.

Now, according to a recent and provocative Accenture survey, it appears that further discussions may be worthwhile: Namely, what happens if (or when) AI software moves from being a guide, to a companion, to a representative of a “friend” worthy of acting on an individual consumer’s behalf when it comes to purchasing decisions.

That’s precisely the question posed by Accenture in their study — titled “Me, my brand and AI: The new world of consumer engagement” —which leads with a few interesting statistics:

  • Generative AI is becoming commonplace, with 72% of those polled (over 18,000 respondents hailing from 14 nations) saying that they regularly utilize related tools.
  • Nearly all (94%) of active GenAI users have or would consider asking these tools for help with personal development goals.
  • Only a slightly lower, yet strong, majority (87%) of respondents indicated that they would query the GenAI tools for social and relationship advice.
  • More than one-third (36%) of those active GenAI users surveyed said they consider the tech a “good friend.”
  • Almost one-in-10 consumers polled (9%) rank GenAI as their No. 1 source of reference when deciding what to purchase.

GenAI Outpacing Social Media and Friends and Family as a Source of Consumer Purchase Recommendations

While it should be mentioned that many metrics quoted involve active GenAI users, a demographic label that not all consumers can currently claim, the numbers presented by the study remain intriguing.

Accenture framed the GenAI journey in commerce as analogous to a trusted “friend,” in three particular modes: as a trusted guide, as a loyal companion, and as one’s “second self.”

In terms of being a guide, the data indicated that GenAI was now second only to the percentage of active users who preferred a physical store visit for recommendations (19%), coming in just below that figure at 18%. Other prominent channels for product recommendations followed in suit, from social media (15%) and friends and family (13%) to search engines (11%) and brand websites or apps (10%).

“One in two users have already informed a purchase decision using GenAI, making it the fastest-growing source for recommendations and advice in the past year,” the study authors wrote.

The crux of the second mode, of AI being a loyal companion, lies in improving AI’s ability to “feel human,” offering an authentic “personality.”

“Consumers distrust AI-generated content lacking authenticity (41%) and personability (45%),” the authors continued.

Agentic GenAI: A ‘Zero-Touch’ Engagement Paradigm?

Likely the most interesting dynamic brought forth by the survey was the notion of agentic AI replacing consumer touchpoints in part, or even in whole.

“Consumers are ready for AI agents to purchase on their behalf, with 75% open to using a trusted AI-powered personal shopper that understands their needs,” the study proclaimed.

The authors indicated that with bot-to-bot transactions and AI agents proactively acting on behalf of users’ purchasing instructions, AI is positioning itself as the ultimate decision-maker in tasks previously led by humans — such as product comparison, checkout, and post-purchase support.

The solution offered up? “Prepare for ‘zero-touch’ engagement,” according to the study authors at Accenture.

Cornerstone strategies suggested included prioritizing experiential and emotional value — giving you an advantage beyond hard metrics like price and product specs and helping you stand out from AI-commoditized transactions — while (somewhat paradoxically) also embracing AI integration wherever possible.

“Brands should strive to build, or integrate into, an AI ecosystem that transcends single-brand silos,” the authors concluded.

Discussion Questions

Will agentic GenAI actually replace human-led buying behavior in a significant way in the years to come? Why or why not? What caveats or headwinds come most quickly to mind?

What obvious issues could arise from widespread usage of agentic AI being empowered to make semi- or fully-autonomous purchasing decisions on behalf of individual consumers? What precautions should be taken to prevent or curtail these risks?

Is it presumptuous to speak of a time when artificial intelligence agents are perceived as more than tools, but as “friends” worthy of being trusted not only with one’s personal details, but also one’s wallet? Or is that day inevitable, at least for a portion of the population?

Poll

12 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Neil Saunders

Some of this is typical tech nonsense. What happens is people take a grain of truth – that AI will play a role in the shopping process – and take it to extremes. In many cases, AI will not replace the consumer because shopping is not merely about the acquisition of goods. It’s about the process, including sociability; it’s about immediate moods and feelings that AI cannot predict; it’s about tangible things like the touch and feel of fabric. Yes, for some basics purchases AI may automate; but that can be done now via subscriptions. What AI will do is facilitate shopping – acting like a personal shopper, helping with search and curation, and forth.

Doug Garnett

We’ve heard all this before. It is fully nonsense. Remember when Alexa was going to be automatically ordering on behalf of the customer? Er. That didn’t end up happening. And with nearly 30 years of my buying history, even Amazon with its incredible technology can’t seem to recommend even a book that I care to buy. We MUST stop reading sci fi books and believing they are the future of retail. This article worries me more because AI should NOT be dominating discussions within retailers. AI is a lovely added algorithm — but it’s only an algorithm. There will be some good to come from it. And given what history reveals about investments in the internet (a far more reliable algorithm), retailers will waste far more money than they ever get from it as long as they keep chasing shiny baubles.

James Tenser

It’s a little too early to know whether personal digital agents will represent consumers in their commerce transactions, but I would not rule it out for the “AI Natives” who are now being born.
The concern with this type of intermediation is the loss of touch that it creates. It’s one thing to let a tool perform a function autonomously, but quite another to rely on that tool to represent you in a relationship.
To make matters worse – or funnier – imagine a scenario where your personal AI agent interacts with an AI agent that represents the interests of the merchant.
This is the stuff of dystopian comedy.

Bradley Cooper
Bradley Cooper

Smart homes and vehicles may be the gateway to Gen AI acting on behalf of the consumer. Think your Home or Vehicle agent may monitor and automatically schedule routine services, or order replacement parts. Smart home shelving may notify your agent to replenish certain designated items when they reach a certain low stock threshold.

In the near term for more nuanced purchases Gen AI will mainly serve to compare and recommend based on the prompts and instructions with the consumer ultimately making the final buying decision.

Scott Norris
Scott Norris
Reply to  Bradley Cooper

But that’s not Gen AI; that’s just a min/max or calendar/mileage-based algorithm. We were doing stuff like this back in the 1990s. Even if it was “watch my usage of pizza flour against projected dough usage for the next 90 days based on historic sales for the period and these leading indicators and order another truckload with 5-day lead time ONLY when storage bin capacity is less than 1,000 pounds” is still a Windows SE kind of job.

Mohamed Amer, PhD

We’ve become accustomed to delegated decision-making in our daily lives—think subscriptions, standing orders, or financial advisors. The Accenture study shows that we have the opportunity to take this to the next level. Daniel Kahneman’s groundbreaking work on System 1 versus System 2 thinking highlights that for low-stakes, repeatable purchases, consumers already navigate choices in a semi-automated manner. Imagine how much easier this could be with AI agents stepping in to reduce cognitive overload in these decisions. However, when it comes to purchases tied to identity expression, social signaling, and experiential value—the very essence of retail—nothing can replace the human touch.

The exciting prospect for AI lies in its ability to change the basic search functions and algorithms into personal curators. These intelligent agents could learn your preferences across various categories, seamlessly connecting your desires with market offerings.

Yet, we must tread carefully due to the principal-agent problem: When an AI agent makes a decision, whose interests does it truly serve—yours, the platform’s, or the highest bidder’s? The path forward for retailers is to develop a hybrid model where AI agents take the lead in preference matching, while consumers maintain control over value-driven and experiential choices. 

David Biernbaum

The use of agentic GenAI can significantly enhance decision-making processes through the analysis of vast amounts of data, but it is unlikely to completely replace human-driven purchasing behavior.In buying decisions, intuition and emotions play a significant role, often influencing decisions that cannot be predicted solely based on data.Moreover, factors such as the lack of trust in artificial intelligence, privacy concerns, and the need for personalization present barriers to full adoption.

Lisa Goller
Lisa Goller

Agentic AI adoption will take time but early adopters will likely start with reordering items they already know and trust. Delegating product discovery through purchase to AI is a much bigger leap.

Brian Numainville

While there is certainly a threshold, at least currently, for what someone might delegate to an AI agent. The higher the price or accuracy needed, the more challenging of a hurdle it will be to “trust” AI agents. And then there is the issue of trusting an AI agent with your payment information. That said, this will continue to grow – I’d love to have an agent in place for routine purchases, low risk purchases, or personalized discovery of new products I might like that match up against my interests or prior purchases. This isn’t as far fetched as some think and not everyone wants the “joy of discovery” for every product purchase.

Scott Benedict
Scott Benedict

A key element in answering this question is related to the concept of “trust”. Does the retailer have the needed degree of trust with consumers to allow them to use this technology on their behalf? Does the technology make their life easier or provide any unneeded stress? In short, if the benefit of Agentic AI makes life easier for a busy consumer, and does not violate their trust, there’s a chance for this technology to gain widespread adoption. Those retailers that have a reputation for being trustworthy in the eyes of consumers, and can demonstrate the benefit of using the technology in context, will earn the opportunity to use it with their customers and gain share as a result.

Kieran Cloonan
Kieran Cloonan

Sure. On some things, especially commodity or replacement items, Agentic AI will very likely have a part to play in the consumer journey. But I feel like a re-read of the old _The Cluetrain Manifesto_ from 1999 is in order. Truth and authenticity in the marketplace are still of paramount importance to customers. With so many AI bots pushing fake voices into our daily lives, consumers are constantly judging trust and authenticity in their interactions with companies. It’s going to be critical for retailers to understand when and where their customer is looking for that human element and take care not to let technology or AI get in the way of that.

Shannon Wu-Lebron
Shannon Wu-Lebron

Will agentic GenAI actually replace human-led buying behavior in a significant way in the years to come? My answer (as most things in life) is: It depends.

I believe it will have the biggest impact in routine, need-based, replenishment type of shopping. For example, when household items need to be refilled. Another good use case is occasion/event based shopping, e.g. purchase a list of things I need for a 3-day camping trip or get me a complete outfit for a summer family reunion in Vermont.

When it comes to other type of shopping that requires high considerations and human decision making, AI won’t be as useful to act independently. Think about the time you bought a car, a TV or a pair of new shoes, it is difficult to codify the rationale that led to the final decision. Humans are infinitely complex in the way that sometimes we don’t even know why we bought something except that it just “feels right”.

BrainTrust

"The higher the price or accuracy needed, the more challenging a hurdle it will be to “trust” AI agents…then there is the issue of trusting an AI agent with your payment info."
Avatar of Brian Numainville

Brian Numainville

Principal, The Feedback Group


"In the near term for more nuanced purchases, Gen AI will mainly serve to compare and recommend based on the prompts and instructions…"
Avatar of Bradley Cooper

Bradley Cooper

Associate VP, Technology, SASR Workforce Solutions


"To make matters worse — or funnier — imagine a scenario where your personal AI agent interacts with an AI agent that represents the interests of the merchant."
Avatar of James Tenser

James Tenser

Retail Tech Marketing Strategist | B2B Expert Storytelling™ Guru | President, VSN Media LLC


Recent Discussions

More Discussions