rafapress/Depositphotos.com
August 11, 2025
With AI Search No Longer Deemed ‘Experimental,’ How Can the Tech Shake Remaining Trust Concerns?
With Perplexity already partnering with PayPal to allow for purchases directly in chat, and ChatGPT slated to bring an integrated checkout system online (with partners including Shopify rumored to be on the roster), the “seismic shift” toward AI-driven shopping and brand channels appears to be on the cusp of realization, per CNBC.
“Enabling customers to purchase without leaving the chat will have a significant impact on the sales cycle,” Elizabeth Parkins, professor of practice of business administration and economics at Roanoke College, said. Parkins added that by streamlining the purchase cycle, sales could be accelerated.
“No more time-consuming steps. Customers get what they want faster, with less hassle, and honestly, with less chance of changing their mind,” Parkins concluded.
New Report Suggests Mixed Trust Signals, But Yext VP Indicates AI Has ‘Earned Trust’ and Is ‘No Longer Experimental’
However, according to a recent Yext report titled “The Rise of AI Search Archetypes,” there remain some lingering concerns — particularly surround trust — tied to the proliferation of the relatively new tech.
Among the major findings presented by the study, which surveyed “2,237 consumers across the U.S., U.K., France and Germany who recently used conversational or voice AI to find information online”:
- AI has earned some degree of “trust” in brand search: Nearly two-thirds (62%) of respondents indicated they they trust AI models to help navigate brand decisions, approximately matching trust levels tied to traditional search methods.
- AI search adoption grows among consumers: Almost half (43%) use tools such as ChatGPT and Gemini on a daily basis, signaling high rates of adoption and integration among contemporary consumers.
- AI not entirely trustworthy as a sole source, however: Only a very slim cohort (10%) trust the first result offered up by AI models. Of those surveyed, 48% conduct due diligence to cross-check AI results against traditional search and other methodologies.
Finally, a generational divide is also in evidence.
Zoomers, millennials, and members of Gen X all engage with AI tools for a variety of reasons (primarily brainstorming, quick insights, and general knowledge queries, respectively), but baby boomers are seemingly much more averse to AI adoption. According to Yext data, 26% of boomers do not use AI at all for common tasks, with the majority preferring search engines for fact-finding (80%) and navigation (68%).
However, at least one exec with Yext seems enthusiastic about the current level of trust afforded to AI search tools, particularly around consumer product discovery.
“AI search tools are no longer experimental, said Mark Kabana, VP of data innovation at Yext.
“They’ve earned trust by consumers across a wide range of use cases, especially brand discovery, where accuracy and clarity matter. Consumers are using these tools to make real decisions. If your data isn’t structured, consistent, and optimized for how modern platforms interpret it, your brand risks becoming invisible to entire segments of your audience,” Kabana added.
Discussion Questions
Do you believe the research presented is accurate, and what is your take on the current state of consumer trust in AI shopping as an emerging process?
What steps can retailers / AI platform partners take to ensure that: A) Consumers will actually find their product listed prominently within AI shopping channels, and B) That consumers will trust those channels enough to purchase goods through them?
Is it likely that older consumers will eventually sign on with AI shopping in the next few years?
Poll
Loading …BrainTrust
Frank Margolis
Executive Director, Growth Marketing & Business Development, Toshiba Global Commerce Solutions
Cathy Hotka
Principal, Cathy Hotka & Associates
Mohamed Amer, PhD
CEO & Strategic Board Advisor, Strategy Doctor
Recent Discussions
More Discussions
Would Recording In-Store Associates Help as a Training Tool?
Ethosphere is banking on AI-driven recordings of in-store associate practices as the next big thing in…Some Retailers Are Trimming Holiday Sales Discounts: Is This a Dangerous Play?
Some retailers are pulling back from deep discounts this holiday season. Is this a dangerous play?Has Kroger Discovered the Limits of Automated Warehouses?
Kroger is winding down its relationship with Ocado, citing cost overruns and a reorganization of its…
The trust issues are not easily resolved. But transparency is essential – clearly stating that results are generated by AI and providing basic insight into how the technology works can help. Ultimately, however, the real drivers of trust are accuracy and relevance. If the information delivered is honest and genuinely useful to consumers, confidence will grow. Unfortunately, we are not there yet. Even basic AI-generated summaries on Google still contain factual errors and hallucinations that undermine credibility. And that’s before we get to the issue of results being driven by commercial considerations versus consumer relevancy.
Search results have always been biased, from those that pay for placement to those that can afford an SEO-expert. All AI, not just search results, needs citation (as Gemini does) as a means of validation, as well as a ‘generated by AI’ label.
Great comment, Frank. It’s not like search results have been neutral at any recent point in time!
“Trust. Yeah, that can be negotiated.”
Well, it’s still not giving great answers. A friend asked Chatgpt to name all the rep side to from Herbert Hoover until today. It was completely confused in one, flipped Biden and Trump in one I tried and when I told it it was wrong it said “sorry, I didn’t QC the answer properly.” Doesnt exactly engender trust.
so if it works better then I’m in. Ironically, before this test, I was using it more frequently than Google. Including for medical advice. It lost a lot of trust with me
Retail will continue to explore AI, but counsels general are going to need to hear about more correct results before OKing AI for customer-facing applications.
It has been my concern with AI technology from the beginning that it will fall into the hands of the unethical world of mostly invisible rats who will abuse it for greed, trickery, and wrongdoings. Before AI is correctly regulated, law-enforced, and carefully monitored, if that is even possible, this will be an issue in the wild west of artificial intelligence.
In the meantime, let’s talk about consumer trust and confidence.
Ethical AI use in retail is crucial to ensure that consumer data is handled responsibly and that AI algorithms make unbiased decisions. By adhering to ethical standards, retailers can prevent misuse of personal information and build a reputation for integrity.
This approach not only fosters consumer trust but also contributes to a fair and inclusive shopping environment.
By implementing rigorous verification processes for listed products, retailers and AI platform partners can enhance consumer confidence. In order to build trust, they should also provide transparent reviews and ratings from real customers.
Furthermore, by providing robust customer support and easy return policies, AI shopping channels can provide consumers with a sense of security about their purchases.
Yext’s data contradiction shows that consumers are using AI as a starting point, which is how they historically approached traditional search results. Trust remains fragile as ChatGPT and others continue to rely on training wheels. Building AI shopping trust operates on the product and purchase sides. Product visibility (discovery) requires investing in schema markup, providing detailed product attributes, and creating anticipatory content to address AI queries. For purchases, we need concrete trust signals that combine transparency, verification, and accountability. Regarding older consumers, AI will need to evolve to meet their preferences rather than expecting them to adapt.
Retailers that pursue trust through AI Agents are missing the point; AI is a trust amplifier, not a trust replacement. The transition from experimental to trusted requires better business models that align AI incentives with consumer outcomes, not conversion metrics.
With the right prompting, AI can serve up some pretty amazing content. Content and context that might not have been thought of originally. It can expand the range of ideas and considerations. AI can also be flat out wrong. It can hallucinate. I’m reminded of the phrase, “Trust, but verify.” With AI, I would say, “Don’t trust. Verify.”
AI can absolutely hallucinate and provide inaccurate answers. So can content created by people. I’ve used AI for a number of purchase decisions, including deep research (Google’s is better IMO), and found the results to be quite good (and accurate when I verified) IF I’ve done a good job of asking the right questions.
AI shopping has potential to change how people shop online, but trust will be the deciding factor. If shoppers don’t understand why they’re seeing certain results or doubt the accuracy, they won’t complete the purchase in the chat.
There is no doubt that AI is taking over the typical Google search. Trust will always be an issue. Even before AI’s rise in popularity over the past two and a half years, consumers often questioned the accuracy of ratings, descriptions, and other information. AI doesn’t look for websites that are optimized for a Google search. It looks for the best information to help educate the customer (in addition to helping them find the retailer to buy from). The way we’ve written advertising and marketing copy is changing. If you don’t already recognize that, you’re already behind.
Deliver value. Reduce friction. Deliver as promised. Trust will follow. Even late adopters are eventual adopters.
Of course, older consumers will adopt AI shopping. Can we all agree how hard it is for older consumers to search for products and prices across sites (or even in a single ecosystem) and find the best? Current shopping is search result overload, and now riddled with interruptive retail media ads. Just like Google Search eroding, shopping behavior will go to the least friction that gets you to your answer faster – especially for older audiences that didn’t grow up native to digital commerce. Their kids just need to show them how…