AI Answer Liability, AI-Referred Shoppers and How to Pick AEO Tools

0:01Hey everyone, welcome to our episode 7 of marketing to agents podcast. I have with me here Lalit. Hey Lalit, great to have you here.

0:12Hey Ashwani, always a pleasure.

0:15Fantastic. So Lith, we saw some interesting news again of course with a lot of things happening in the market. This week as usual was some key updates, some interesting changes in what we are seeing in the market. So happy to have you here and let's get into it right away. Okay. To start off like first I think what there was an interesting u news coming out of Germany. So basically Germany's Zach ruled Google AI overview and perplexity are content providers. So to just give you a quick context before that so till now all the all the Google news for example was treated as an aggregator a passive aggregator. So they were never a content provider, never a news creator or news capturing. They were more of a just an aggregator. They were never treated as publisher. But there is this new ruling out of Germany where they are suggesting that any AI overview or even perplexity answer would be treated as a content as a publisher style media li law liability rather than just a aggregator. So seems to be something which might have a bigger impact across the board across the industry. what are your

1:33Thoughts? What do you think about this ruling? Yeah, I think this is very significant and this is high chances of trickling down to the entire EU and once that happens I think this will be a significant market that expects you to you know own up you know more than more than what what people are used to owning up right especially with respect to the these AI aggregators it is it is traditionally being seen as aggregators I think I can understand the the rational behind behind this ruling because we have started to treat AI models and AI agents as digital employees, right? And and the work done by a digital employee cannot be you know sort of you know the the company employing a digital employee cannot distance themselves from the digital employees work. So I think in that spirit or in in that lens the ruling actually is in the right direction. but definitely it opens a can of worms because you know ultimately it is not true human work. right so if this if this becomes a reality I think

2:50Compliance basically compliance on all these norms and you know this might become like a big thing for almost all publishers who are using AI AI to summarize right imagine let's say a small startup which is using AI to summarize just emails or let's say a business dashboard right so the the scope may be today AI search but I think it sets the direction to include a lot of AI related summarization that is happening almost on a rampant basis in so many digital products I think all of it can be under this perview and will just increase the cost of compliance

3:35100% so while as you said I think while on one side this increases the owners and responsibility on the LLMs and the LM provider providers. I believe this is also a welcome relief for the consumers, the people who are consuming that information because today there is no absolutely no liability today on the engine itself on an L&M itself of how well regulated it is, how correct their information is. There is absolutely no control on that. So even if AI overview is reporting false information finally the onus falls on the brand or on the publisher that okay publisher may have provided the wrong information or AI the the LLM would have picked up something wrong context of that publisher. I think in my view it is also a slightly more u it's good to put the LLMs also on a pedestal where they are treated at par because they because they the impact that they have because the control that they have today they should be responsible for some in some parts to adhere to certain level of accuracy certain level of control in what is being reported again LLMs of course the whole AI being a blackbox they will have to find ways to regulate it or not

4:52Regularly but at least in a way at least make sure it is accurate. Make sure the information provided is accurate. but yeah like you said it would of course increase u increase the amount of surveillance and amount of information required to then keep them in check. Yeah and I think you're absolutely right from a consumer point of view this is this is actually kind of a relief right because I used to you know at one side my reliance on AI lms and perplexity kind of products is increasing and the moment I see let's say a disclaimer that AI can be wrong right it used to just generate anxiety in me right because I'm probably you know asking something very crucial u and then that this disclaimer that this AI could be wrong would would just does not would not sit right with me. And I think rulings like this would force these labs which have by the way a lot of budgets right a lot of talent out of budget they should be able to invest more into the observability and you know guard rails and should sort of ensure and assure the user that whatever they are basically you know composing as an answer it's grounded in true facts and it is no longer something that will make mistakes right so this this is a welcome change

6:10And I think consumers would could really appreciate that level of scrutiny and preparedness.

6:16Absolutely. And and like you said, I think this could set a good precedent not just in Europe. I think there could be many many geographies which might follow suit because this does put like at least some responsibility on the LLM providers as well and just because they control so much share of voice mind share I think it we'll see this happening I think across the board across multiple geographies hopefully. Yeah. And one second order effect Ashwony of this would be that the liability insurance for AI decisions, right? That I that actually might become like a real market, right? so interesting maybe we'll have a discussions on this sometime.

6:58Yeah. Yeah. This an interesting case in point as you mentioned about this is that there I've Air Canada for example I I heard about that at least a couple of years back. they were one of the first to adopt an enterprise chat interface which was using LLMs to provide answers and then at some level I think in one of the queries it provided a wrong answer which led to a traveler's booking getting cancelled or something happened and then they faced now who's to be blamed was it the LLM was it air Canada of course at this point Air Canada had to take the responsibility and then I think there was some compensation given to that particular traveler But you're right, I think this would definitely put of course the onus on the right person like whether it's the LLM provider, whether it's the enterprise and then at the same time like you said I think the liability has to be covered and who covers that would become critical as well. Perfect. all right with that with that moving on to the next next news and this is coming from Google itself. So Google merchant center that pilot goes live in US merchant accounts. So like anything you would like to talk more about it?

8:18Yeah. So I think Google always have had a significant role to play in the shopping discovery for products right and I think they have aspirations to play a much bigger role and to some extent also you know get the transaction done through this through a lot of the latest advancements like the Google universal universal card that we discussed in one of our previous episodes and also the universal commerce protocol right so I think what this allows like this latest development which is the AI performance insights. So if you are a shopping merchant whether you have a site on Shopify or Magento or even WordPress right Google is anyway able to crawl your catalog and now it is going to also expose you know AI performance insights like the the catalog completion score or how can you optimize your catalog description or the title of the product so that it performs better on search. it sort of you know gets into the right search queries and gets in front of the right customers. So I would say this is a this is a great change you know especially for B2C or T2C commerce brands this is this is going to be very helpful and obviously

9:37Push puts Google front and center of this agent commerce evolution. Absolutely. Absolutely. And related to commerce, there's another news which came in this week which was from the prime date prime day 2026. We saw that just finished recently and some interesting data coming out of there. the data is that AI referred shoppers converted 40% better than other channels. So if for example on Amazon there are certain products which were referred by AI by AI and those shoppers came to buy there was 40% better converting than traffic coming from other sources. couple of interesting insights out of this. This of course is a huge jump over last year u and also interestingly AI is not just a discovery thing right now. They're not discovering products using AI. This is AI is actually driving conversions. So they do all the discovery and once they come to the site, they're actually converting faster. So very interesting data. And since we're talking about the merchants and we're talking about the catalog,

10:54Looks like AI is not just converting AI is just not helping discover and drive the user to a right decision as well. And not just on the on B2B side looks like consumer data also is not reflecting that and Amazon being the largest e-commerce I think that is some level of data to show that yeah in one of our previous episodes we discussed that there will be two surface areas for information consumption one would be browse the other would be search right so browse is is still going to be let's say surface areas like Tik Tok Instagram and maybe the Google discovery feed and maybe a few others, you know, something totally new. And this the this data that Prime Amazon team has shared, it actually reinforces that same thing that we discussed where the brand or the true demand generation, right? happens actually on another surface area which is upstream of AI search and AI search essentially is a is a is a surface area to know more about the brand get more let's say confidence in let's say making a final choice or a short list and I think in future both these will both

12:12These surface areas will continue to be very important brand building will continue to be super important and our Airmeet suite of products actually help you know B2B brands especially to to sort of have a strong brand building effort going and our AirPulse suite of product basically helps your B2B brands again to to make sure that they're performing really well when it comes to the conversion side of you know the the surface area this would be it would be interesting and maybe I'll look for that information as well it would interesting as to how much percentage of traffic was from AI was AI referred traffic because as we know that there is a huge search budget which is controlled by Amazon as well. So a lot of merchants actually advertise on Amazon itself. Now, if the AI traffic is already coming with a clear conversion intent, that means they've already done the discovery. They've already decided what they're going to buy. They're just coming to Amazon probably to just do the transaction. So, if this traffic starts becoming a material percentage of traffic coming to Amazon, then at some point it might start hitting Amazon's

13:31Maybe the re the ad revenue which they are actually generating from ads running on Amazon. I'm just thinking like maybe way maybe massive user behavior change once people actually use AI for research and not really go for discovery or not really go for category searches on Amazon eventually. But yeah, I would say if that shift starts happening, I think there would be some impact we might see on Amazon's ad revenue as well.

13:58Absolutely. Yeah. I think just yesterday I I purchased something on Amazon and I think even the best performing merchant had to advertise, right? So it was if you we just maybe pass through the sponsored panel

14:19The merchant that I you know so I basically made two clicks. one was on you know on the sponsored panel I just looked through like you know the merchant with the highest rating and then when I came you know reached the organic panel of search results that same this the merchant that I clicked on was already listed on top. So despite being the best merchant you still have to continue to you know push money on paid discovery sponsor discovery. So this is this is kind of like you know maybe maybe it's a it's a necessary evil for brands to continue to outperform the competition but I think with AI discovery u if AI continues to play an important role in discovery also I think maybe off you know on a on a surface area which is off the Amazon products I think over there maybe brands might have a relief where the com customer would evaluated it on charge or perplexity and then a single click to their Amazon listing page would would just you know have them save a lot of lot of advertising dollars. So AEO can so especially B2C brands can look at AEO as

15:38As a way to sort of bring down their advertising spend on Amazon. Like if you're if as a brand if you're spending too much on discovery for on for on Amazon itself. So if you can divert some of the budget to AEO I think you might actually end up getting better results.

15:57That's very true. I think that has a strong implication. That's that makes a strong case for brands to actually start investing more in AEO AEO search and ensuring that their rankings on AI search their mentions their citations are actually top-notch and tracked aggressively. Perfect. well that brings us to the next news or next piece that we are going to discuss which is around around the same topic that we've just talked about which is AI visibility. Now if I'm a brand like there are there there is a plethora of AI visibility tools available in the market today right and it's a highly fragmented place of course u we have AirPulse as well and there are other other tools as well now as a brand what's what is the insight that we are seeing in the market so as a brand there are there is we are seeing a clear distinction between two set of players there There are clearly two set of players which are available in the market now. One set of players is which is purely just focused on just tracking monitoring prompts tracking visibility and that's it. Probably some amount of competitive

17:14Benchmarking as well. And then there is a separate set of products which are actually going much deeper and not just limiting themselves to only monitoring but going further to even detect and first of all of course monitor detect gaps and then eventually remediate itself also. So probably closing the loop with the right content ensuring that the content is crawled is is properly indexed and eventually gets the gets started starts getting featured in AI searches or get starts getting those mentions. So there seems to be a clear distinction now getting build in the market. well where do you think brands should focus more or what kind of solutions are better for brands to use? what's the market how does this market landscape how should a brand understand this market landscape?

18:09Yes, I think as you rightly said I think the the monitoring aspect and the remediation aspect. So there [clears throat] are tools which are just doing monitoring and that by the way has been largely commoditized by sort of you know giving your cloud code or a cowwork a set of skills and then having a routine right you set a schedule task to run it every Friday or every Saturday and then you get the report. So although we can debate about this extensiveness and as we have already discovered through the podcast that you know the monitoring itself is getting multi-dimensional with not just the model but also modes in different models right so it's it's also tricky and what kind of problems do you monitor on is is is going to be the key thing right and as these engines are becoming more and more personalized I think people are not just it's it's no longer relevant to just monitor the category level query but you have to go long tail right so monitoring itself is not something that is trivial but having said that it is still getting a large set of players which are jumping in because they believe it is very easy and so the noise on that side of the market is actually very high right but still players which are very detailed can

19:26Stand out on the on the other side there are let's say somewhere an end to-end promise where the the platforms will detect what is wrong, where where the optimization should be made with respect to the right set of queries and then they also present a proper action plan and to some extent agentic ways to automate the remediation activities also and most importantly bring back a feedback loop after a certain remediation has been executed. What actually happened on the monitoring metric, right? Did it improve? Did it did it was neutral? So the the the the other side of the bucket which is more end to end kind of solution I would say the key value is the feedback loop and the quality of the recommendations right with the which the feedback loop filters out right so you're not just being bombarded with let's say your stock recommendations which are floating around it those tools can actually surface data back data backed and you know battle hardened or battleproven techniques which actually can drive results. So at AirPulse right so we we are in the category where we not just monitor we also help brands with a detailed action plan because websites

20:44Are also a very very fragmented landscape right so not many websites are still on WordPress or on let's say on legacy players like fix some are also on let's say new age platforms like you know your lovable etc right so remediation takes time the remediation would be fragmented also right the the strategy to remediate so we do one extra step after monitoring is that we give you a detailed action plan which brands with let's say team members with relevant skills can go execute themselves and we are also adding more and more automation on the remediation side. So we recently launched a WordPress plug-in. The WordPress plug-in allows you to sort of take action on the audit actions the recommended actions that the the audit step is generated and it can essentially put the whole AU optimization or autopilot for our customers who are using WordPress. We now are also launching our own CMS right so your entire content publishing can be managed by AirPulse. So I think for brands it's a clear sign right? If you are if you if you are early in your journey obviously start with monitoring tools but if AU becomes super critical right if AO is your is your strategic you know territory to win right then I

22:02Think best would be to sit with detection to remediation kind of platforms like Apulse fantastic way to open and good demarcation as well. now within so of course AirPulse is a tool which goes from monitoring right from monitoring to eventually remediation as well and helping close the loop from tracking to finding the gaps and then eventually filling those gaps and then even eventually looking that back into the tracking. Yes. now like if I'm a if if there is a brand what should be and as they are making up a decision that okay now they now is the right time for them to switch from let's say just pure monitoring to now okay now they need to start building the right content now they need to start influencing the right right places on internet let's say Reddit kora to be able to get the right citations for their content or mentions u what do you think that trigger should work like like is it just from when they are looking to start growing or when they're looking to start improving their content what do you

23:19Think that trigger looks like when airpulse a product like airpulse is the right product for them for a brand so I think if someone is just trying to maintain a bare like a bare minimum web presence then probably platforms like airpulse are not are not really right for them. but if as I said you know the winning the go landscape is important for them which which I think in every growing category what whatever categories are growing and the market is expanding in those categories I think they should prioritize answer engine optimization from the day one right because there is also a time value in this domain and and there there there is an aspect of compounding benefits right So there is a good amount of weightage that is given to roughly 40% is given to let's say earned citations like earned mentions on third party media which also includes your review websites and customer testimonials on LinkedIn and there's good weightage on let's say paid partnerships with the ecosystem where you can do let's say a guest post or a joint study with someone right so the

24:37Earlier the the early you get into the game right earlier you get into the the advantage comes early and then that can compound right. So if you have more customers because of a better AO in the first phase itself, right, those customers when they write testimonials, when they write reviews, right, it just compounds. So it sort of reinforces your leadership position you know, in in a very strong way. So I would say if at all there is a there's a growth, you know, expectation from the AI search channel, right? So typically all businesses which have some sort of a global distribution let's say a digital distribution or or [snorts] a digital way to acquire customer they must invest in AU and as early as possible.

25:20Perfect. That's a that's a fantastic suggestion and for brands who are listening I think that's a great that's a great segue for for you to take a decision as and see this investment as a priority investment on building your AI presence and building your presence in AI search and ensuring that you're getting mentioned for the right prompts. Well, with that I think we are we have covered some of the key insights from this week. anything top of your mind other than other than the news that we've seen till now?

25:58Yeah. So I think the AI max campaign toggle in Google's Google ads right that I think we can discuss. So so Google had launched the AI max toggle which essentially sort of you know unleashes AI optimization for better queries to target. you know the AI max the AI max the AI max mode enables Google to determine what extra queries what extra landing pages what extra content optimization on your ads can be possible right which was launched in a way that brands were experimenting with because they had no controls on it now what Google team has done is that they have introduced brand controls where they can explicitly say that you know to don't get into territory of keywords which which has let's say a competitor mention or don't include this set of let's say landing pages from my website or on the copy you know you know sort of have these guardrails and they can also sort of steer the AI max optimization by giving like a proper prompt. So this is again is a is a is is a good development because AI the AI mode and AI AI max sorry Google AI is anyway increasing its

27:15Market share it's increasing its relevance and still a lot of AI advertising still is getting funneled through the AI sorry the Google chats right so so I would say that that again is something that we we should discuss

27:33I feel yeah that's a good point And u you're right. I think till till the user behavior remains extremely in a flux I think we will keep seeing this p I think it'll be hard to land on what will really work. So for example a lot of today the or today the page searches of course below AI overview but then now within AI max search property I think those searches are coming next to AI overview. So I think Google is trying to figure out what where will this user behavior land and how valuable and which properties get the right amount of clicks so that Google's search revenue keeps on increasing. So I think eventually I think these are still early experiments. Google is also trying to figure out where and what are users clicking. Maybe they'll try something with the citations. Maybe there might be eventually a property within mentions also that there could be mentions also which you can advertise eventually. I I don't know. I'm just again hypothesizing that you know what all is possible like the other day I saw ads in chat GPD I'm sure a lot of people are seeing that those ads those ads today are completely out of context right so if you're searching for something around shoes the ad could be about furniture so out of completely out of context just maybe

28:50Because they don't have enough advertisers they have not built enough context today but I think eventually but those are not purely not search they're just u in a way they are more of you know display ads where they're just showing another relevant category brand there. But yeah, hopefully I think all of these I think are early experiments and finally once the user behavior is slightly more firm, I think that's when we'll know what some of these inventories really work out as.

29:22Absolutely.

29:24Yeah.

29:27Perfect. let's so I think this was again like always this was very insightful lot of learnings there and again to the listeners feel free to share any feedback any thoughts any questions that you have and we are happy to cover them in our subsequent episodes. All right then u right thank you so much for being here. Thank you everyone. Thanks.

29:52Thanks. Bye.