0:00 Welcome to today's session, and thank you all for joining defining authenticity, trust, and responsibility in an AI reality. 0:09 I'm Alex Kirk. I lead the insights analytics and measurement team here at Bazaarvoice. And the insights function of our team is focused on keeping a pulse of consumer behavior and sentiment. 0:21 And as I AI tools start to have more of a prevalence in the shopper journey, we've been asking shoppers directly about their actions, adoptions, and overall perception. 0:31 For today's topic, over the past few months, we've run three surveys to over six thousand nine hundred shoppers to understand their thoughts and behaviors as it relates to LLMs, AI generated product recommendations, and or summaries. 0:46 And with our most recent round of research, which we'll be sharing today, we surveyed thirteen hundred shoppers who we broke out into specific cohorts to glean insights about their usage of AI tools to assist in their review creation, and then the perception of reviews when consumers know an AI tool was used in some capacity. 1:07 So how do shoppers feel about this intersection between AI and the online product review ecosystem? Well, the findings suggest that while shoppers may not be completely trusting of AI's role, the core issue isn't simply its usage, but how it's applied and communicated. 1:26 As AI becomes more common in content creation, consumers are making it clear that authenticity, transparency, and preserving a reviewer's personal voice and valid personal product experience remains essential to maintain that trust in reviews. 1:41 Before we jump into the latest round of research, I did wanna take a moment to speak to the adoption and prevalence of generative discovery. 1:49 You may have heard or read some of these insights from our previous research in previous sessions. However, it's important to connect the dots here to show some of the macro trends. 1:59 As we all know, shoppers have traditionally leaned on search engines like Google where keywords lead to top results. Now as consumers are able to express intent naturally through generative AI tools, we're seeing that three in four shoppers are using these tools for product recommendations. 2:17 Search is evolving into generative engine optimization or GEO, which is much easier to say. And while over half use generative AI tools as a supplement to their research, nearly twenty percent have abandoned traditional search entirely in favor of these AI tools. 2:35 Shoppers are entering search in discovery mode without a clear destination. They're exploring, comparing, and actively looking for recommendations. 2:44 And in order to be discovered, your brand must show up where these tools are looking. We also know that shoppers do not trust AI summaries blindly. 2:52 They trust the human voices powering them. So this is a really important call out. This is new terrain, and when there is a lack of visibility of what information is ultimately sourcing these recommendations, we do see an increase in skepticism. 3:08 However, when we ask shoppers what sources would alleviate the skepticism and build trust, close to two and three agreed that the AI recommendations would have more credibility if sourced by reviews and photos from real people. 3:24 And if you take a look at this slide and the breakdown of these responses, we could imply that user generated content from fellow shoppers could be up to seven times more influential than the brand provided information. 3:37 So to be discovered, it's not enough. You must also be trusted. Even with an AI recommendation, nine in ten shoppers require additional validation from those real customers before they buy. 3:51 So what does that mean? Well, following an AI generated recommendation, a consumer is still going to go to the PDP to read and see user generated content. 4:01 They need that additional validation to really prompt them and push them to purchase. And as you can see here on this slide, this importance is consistently high regardless of the industry. 4:13 So you could say that authentic human voices are serving as that nonnegotiable anchor of the PDP. Okay. So knowing that authentic user generated content gives consumers confidence, let's see how savvy all of our AI barometers are. 4:28 Let's take a look at these two reviews for a moisturizer. Take about twenty seconds to read them, and then tell us in the poll section which one was written by a consumer. 4:37 Alright. I'm seeing some responses come in. Alright. So if we're looking at these, it seems like people are very savvy, so good work, everyone. 4:47 Review b certainly had the majority there. And if we compare the two so looking at review a, this was written entirely by AI with a single prompt. 4:57 So write a five star review for a moisturizer. There's no personal product experience or descriptors. And then on the flip side, review b has mentions of dry skin. 5:08 They speak to how the pump is stiff. They talk about how it's impacting their foundation and the duration of use around three weeks. 5:16 So all those details and nuances that you'd want to know as a shopper. And these details can only come from someone who's actually bought and used the product. 5:25 According to our latest research, sixteen percent of consumers feel very confident that they could distinguish the user generated content from AI led. 5:35 So you all over index there. Great work. Knowing that sixteen percent of consumers feel like they can identify human written versus AI written reviews, how many consumers are now using AI in their review creation, and how are they using AI? 5:50 And there's an emphasis on the how, which is really important here. While only four percent of review providers are using AI every time they draft a review, Nineteen percent said they are doing so sparingly. 6:04 So one in four are using AI in some capacity for assistance when drafting their reviews. Let's break this down. 6:13 Of those using AI to assist their review creation, we do see that close to one and two are using tools provided on sites that give inspirational suggestions on topics. 6:25 And then what we also see here is that over a third are going directly to LLMs for that support. We were curious about the consumers using LLMs directly. 6:35 What are the prompts and inputs that they're using? Are they using AI to ghostwrite or copy edit? So we were very pleased to learn that eighty three percent of those who are going directly to the LLMs to draft reviews were drafting the reviews themselves, but using AI to improve the grammar, the length, and to overall make some adjustments to what they had already written. 6:59 So copy editing, not necessarily doing the ghost writing that we we've the next chapter two choices of this question included providing their own bullets about their specific experience or using the specific words that the product made them feel. 7:15 We then wanted to understand the perception of consumers' heart within reading these reviews. How do they feel if they knew AI assisted in the review creation? 7:25 Well, over half agreed that if a review was given a prompt with the unique review experience descriptors and a star rating, then it would still keep that review authentic. 7:36 And speaking of review perception, we asked reviewers who use AI tools to know what their AI x are. What are their frustrations? 7:44 What are their red flags, their callouts? This was a select all questions question, and the responses for each choice were high. The number one ick was the overall the over the top phrase that an AI generated review provides. 7:59 This was followed by the tone feeling too robotic and then the repetition with certain phrases used within the review. We also see that losing my voice is a pain point, so whenever removing that genuine and authentic aspect of the consumer's unique experience. 8:18 And then for those of us who are frequent users of AI and are familiar with hallucinations, this is also a pain point, especially when you know the product and its unique attributes. 8:29 And lastly, the review can lose a lack of nuance that a consumer can provide from their specific product experience. This lack of specificity and voice takes away from the peer to peer connection that user generated content provides for other consumers. 8:46 So how can we assist consumers with their review generation and still keep authentic? Well, we can provide them with the inspiration tools to write higher quality reviews by providing AI generated topics as tips and suggestions. 9:02 I think we can all relate. We've all had that moment of writer's block before, and reviews are certainly not immune to that. And our content coach feature is a great example of how you can use AI to help generate reviews authentically. 9:17 The tool doesn't write anything for the shopper. It's simply using AI to suggest unbiased topic ideas tailored to that specific product category for the reviewer to write about. 9:29 So this preserves the reviews authenticity while also increasing the length and richness of reviews to fuel future customer purchases. 9:40 So consumers can use AI as a tool to help refine their thoughts. And as we've heard thus far, they ultimately trust reviews that reflect real experiences and real voices. 9:49 The x can be a pain point for those using AI tools, but there are also signals to suggest that the reviewer might have used AI to ghostwrite versus to copy edit. 10:01 And this contributes to the current skepticism of consumers and their trust barometer with the reviews they're reading. As you can see here, while twenty six percent of consumers agree that they are very or somewhat trusting of AI's role in the overall shopping experience, forty one percent are very or somewhat skeptical. 10:20 You can also see here that over a third of shoppers are neutral. So there's an opportunity here to build that trust, and they're waiting for a reason to trust the content. 10:30 And when we probe further to understand if a consumer considered a review to be authentic if AI was used to write it, close to two and three consumers said no. 10:41 However, if you recall, eighty three percent were not using AI to write reviews entirely rather to edit a first draft review, which started with the consumer's unique descriptors. 10:53 This provides an opportunity for transparency and clarity of their tool usage. And we're seeing this shift everywhere from LinkedIn to Reddit where there's openly flagging that used AI for grammar and flow. 11:09 This transparency is your opportunity. By providing this clarity, you're not just appeasing the skeptics, but you're potentially converting that thirty three percent neutral group and helping them to feel that they are AI informed and therefore AI ready shoppers. 11:25 This transparency and clarity is imperative as when asked if a review should be batched appropriately, the majority of respondents agreed the content should be noted that AI was used in some capacity. 11:36 And there is still opportunity to educate and build trust when it comes to the appropriate usage of AI with review generation. While close to forty percent of consumers said they'd still likely purchase a product when they knew AI supported the review creation, over a third still feel neutral. 11:53 Again, this thirty six percent is a opportunity. By providing clear disclosure, we can break down those silent barriers to purchase. 12:02 We can help take this neutral crowd and help them to be more receptive and also shift to taking what was once an unlikely shopper and making them a confident buyer. 12:14 So education and transparency is so key here. And how can we take this perception be shifted of, like, what AI usage is deemed appropriate yet still maintains authenticity? 12:28 Every review in your ecosystem should ultimately trace back to a verified purchaser with a real product experience, not a product description or a synthesis of other similar products. 12:39 As we heard, AI can't manufacture the specificity and context that comes with an actual experience, and shoppers certainly feel when it's missing. 12:49 We know that transparency leads to trust, so show shoppers that what they're reading has been verified. In addition to the perception of reviews on your PDP and at retailers, we also know that LLMs are prioritizing certain signals, which then leads to AI recommendations, which consumers are using more and more likely to trust when they know they are supported by authentic ratings and reviews. 13:16 So how do you ensure you surface in GEO? AI agents don't just read your PDP. They're looking for specific high weight signals to validate their recommendations. 13:28 So as you can see here on the slide, here are a few. So, these are five signals as part of your GEO visibility checklist. So the first being a verified buyer badge. 13:40 This could be considered a trust anchor that shows that this content has been written by a verified shopper. Number two, reviews embedded with media. 13:51 So reviews with photos paired with a fifty plus word count, that certainly is a great way to alert the AI agents and explain the why behind a purchase. 14:05 Three, specific language. Keep saying this word specificity, and that certainly surfaces. Generic praise can often become invisible, even to the algorithm. 14:17 So those nuances, that personalization, that is certainly going to be a high weight signal. The next, recency. 14:27 So content older than ninety days, three months suggest a stagnant product. So we certainly wanna be not only thinking about the volume of your reviews, but also that recency, making sure you have refreshed and updated reviews. 14:44 And then a realistic sentiment mix. AI is looking for that eighty twenty balance of praise to criticism to prove that the reviews are genuine. 14:53 And if you're interested in our PDP audit prompt, please feel free to check the document section on your screen, and we do have this as a resource for you all. 15:02 So what does disclosure look like in practice? Over half of review say the most authentic AI prompt starts with, I tried this product and liked it because and giving the various descriptors and their unique product experience. 15:21 This matters beyond the PDP. And as Bazaarvoice's CTO, Nick Shifton, puts it, the combination of agentic AI experiences with trusted verified ratings and reviews could unlock a golden age of personalized shopping. 15:39 This can happen only if the content they're drawing on can be trusted, which is why verified reviews are a trust signal for both shoppers and AI agents. 15:50 This is a great example quote to summarize the importance through both lenses, and this is why verification with moderation is ultimately providing those trust signals that matter most right now. 16:02 And you can start this by ensuring that your ratings and reviews platform using strict third party verification protocols, you know, thinking what good looks like in a content ecosystem where reviews are real, unmanipulated, and verified. 16:19 That combination eliminates any sort of hesitation. It can reduce skepticism, and it reinforces trust at exactly the moment a shopper is deciding whether to buy. 16:29 And for scale, Bazaarvoice's intelligent trust mark helps to automate that verification behind the scenes. It's the verification layer that sits behind the badge proving a review came from a real verified purchaser and that it passed authenticity standards. 16:47 And this is what can convert a skeptical shopper into a confident buyer. Alright. So we've made it through to the end of the session. Thank you for your time. We will have a q and a. 16:58 But before we jump into the q q and a, I did want to bring us to the finish line with some key takeaways. Alright. So number one, stay in the know of consumer AI adoption and sentiment trends. 17:12 This is an evolving space. You know? Everyone is kind of learning each day both from our roles as marketers, but also our roles as consumers. 17:23 So it's really important to ask shoppers directly how they're feeling, what the actions that they're taking, how their purchase decisions are evolving. 17:32 And myself and my team are very committed in asking consumers directly and reaching out to kind of understand from both a quantitative and qualitative perspective. 17:44 Two, remember that the power of user generated content is that peer to peer relevance. So I think one of the words I said the most so far in this presentation has been authenticity. 17:56 It's that authenticity, that relatability that is really moving the needle. People relate to other shoppers. They wanna hear their honest feedback, and we always have to remember that that is what is ultimately driving a lot of purchase decisions. 18:14 Providing resources and tools to inspire never dictate robust and nuanced review. So we heard that close to half of consumers are already using AI for tools to inspire how they can lengthen and create more robust, rich reviews. 18:32 But it's important to never dictate and to really remember that at the end of the day, again, we need that that human voice and that relevance. 18:43 Thinking through how you can optimize your r and r strategy to be LLM ready. So we went through some of those, like, five key points to, you know, make sure that you're thinking that through your strategy so that those signals are very apparent to LLMs. 18:57 And in our document section, there is a a prompt to kind of do that GEO audit of your PDP. That's always really crucial to really be thinking through those optimizations. 19:10 And then in addition to thinking about the signals, it's also remembering to ensure your product detail page is rich with UGC. We heard this is the validation for consumers at that point of purchase. 19:22 So even while more are leaning into geo led product recommendations, there is still some skepticism there, and they're still going to the PDP to further validate with UGC. 19:35 And this is, of course, is the point of purchase, and so that is where they're ultimately making that and taking those actions. So, again, continuing to ensure your PDP is rich with UGC. 19:46 And final words before I think we have just over five minutes for a q and a. The findings suggest that while shoppers may not be completely trusting of AI's role, the core issue isn't simply its use, but how it is applied and communicated. 20:01 And as AI becomes more common in content creation, consumers are making it clear that authenticity, transparency, and preserving a reviewer's personal voice and valid product personal product experience truly remains essential to maintaining trust and reviews. 20:18 Alrighty. Well, let me see if we we have some questions in the q and a. Alright. So first one, how does AI assisted content affect our visibility in search engines or discovery tools? 20:33 Great question. So I I think we kind of touched on this, like, throughout too of where the those signals are weighing the most heavily, and it is really leaning into where they're finding that nuance and complexity with the reviewer experience and that robust when there's media embedded, the length of the review. 20:55 So some of those signals that LMs are picking up are are really kind of leaning into that human voice that is getting provided from the consumer directly writing. 21:11 The next question, there are advantages to AI as long as it's used in the correct way. So are there any parameters that Bazaarvoice is implementing to encourage reviewers to write an authentic review and to avoid writing a clear AI written review? 21:27 Yes. So the this is a great question. I I think that's why leaning into the fact that we can all empathize that sometimes it's hard to articulate your thoughts in terms of reviews, and you want to give that robust review and people can suffer from writer's blocks and why tools that can inspire and give ideas or tips or suggestions of, you know, various topics that are are tied to the product to give ideas out to lengthen or to, like, lean into a particular characteristic are so key. 22:01 And gotta I keep saying the word inspiration, but it is truly that because we never want to be too prescriptive or, you know, we we still want to maintain that authenticity. 22:11 And I do think it it it was great to hear that close to half our consumers are using these types of tools to help lengthen their reviews and that anyone who was going directly to the l a LLMs for review creation, it was eighty three percent were just using that to correct their grammar or to edit or to add nuances here and there. 22:32 This is a good one. So at what point does the review officially count as AI assisted? If a shopper just uses AI to make sure they don't sound mean, are we overcomplicating things by demanding a disclosure? 22:46 Yeah. It's, you know, it's a it's an interesting question. Right? Because we like, spell check or, you know, Grammarly, like, there are ways that we've leaned into editing and, like, reviews and just her like, kind of making edits here and there. 23:04 Agree that it could be overcomplicating, but I do think there's still opportunity and education as this is new terrain. 23:12 And, if there are ways to, you know, use verification to say this was written by a verified purchaser, which has been the norm for a while. 23:24 I I think it's that education and transparency that will get people comfortable and alleviate some of the skepticism. Alright. 23:35 Well, that looks like all the the questions for today. Again, I want to thank you for joining this session. 23:43 If you have any additional questions, always feel free to reach out, and our team will continue to ask consumers directly and kind of gather these insights so that we can all feel more informed. 23:55 Thank you very much. Have a good rest of your days.
