Social’s sneaky role in what AI decides to show

Yeah. So, my talk today is going to be about the sneaky uh role that social media platforms play in what AIs decide to show. And on the agenda today, I'm going to answer some questions about how AIS discover and index social media content, how they rank and retrieve it. Um going to talk about the the impact of engagement and and virality on social media posts and AI visibility. And finally, how we can measure and improve that impact. And we're gonna go over the new type of collaboration that's becoming more and more important between two teams, social and SEO. And I'm going to show you the actual workflow that we have implemented this quarter at Plannable. So, nothing groundbreaking here, but uh just a quick refresher that over a decade ago when I actually started in SEO, Google was pretty much synonymous with search, right? If you had that nailed down, if you if you

ranked well in Google, you you were golden. But fast forward to today, search has really fragmented and users are no longer just going to the traditional search engine. They're going to social media platforms to to search for their queries. And they're also not um any longer restricted to like four or five carefully selected words to describe their uh their issue. They actually now have the whole entire space of chatbots at their disposal to cram pain points and particular criteria and specifics in their prompts and they can rest assured that every time chat bots are going to, you know, answer super thoroughly. And more than that, they're just going to prompt another question to to keep users in the chatbot and engaged and hooked, which can easily lead to like zero clicks for for websites. If we look at the latest SE ranking data about how users are finding websites nowadays,

we can get an idea about how organic traffic is fairing. It's still getting the the lion share sitting comfortably at 47% but we are seeing a downward trend for it. From the beginning of the year till till now it's lost almost 4%. So it's definitely sliding and we can only guess in in whose favor. But yeah, social media traffic is at around 10% which means that one in in 10 people are arriving on our sites from a social media platform. Um and it's starting to be more than than just traffic with social media, right? It's it's about influence. It's about creating content and delivering it where our users are consuming this type of content and it's also about shaping LLM's uh perception of the brand, the company or the product. And finally, AI traffic doesn't seem to have a huge share, but don't

don't let that uh small percentage fool you because we are seeing this share of traffic rocketing and I think it's it's going to surprise all of us when it's just going to keep doubling over the following months. So um we looked at um organic uh search results and um again no surprise here but Google is surfacing more and more social media platforms in in its SERs. Um actually 50% of um um results now feature at least one social media platform. Um and that includes you know the usual suspects uh Reddit, Instagram, Facebook. Um SE ranking actually discovered that um um in top 10 results from Google uh um 37% of the times uh Reddit will be mentioned uh and it's closely followed by YouTube which has a 20% uh share um and then followed

by Kora with 8%. So um this is the the proof that we're all looking for. Um social media platforms are super present in SERs. They are being surfaced by by Google because they know that this is exactly what the users are kind of pivoting towards. You know they want that authenticity that comes from social media. They want to hear from peers and they want to um learn from people who have come across the same pain points that they have. So, we've looked at how social posts pop up in in organic, but how are they showing up in um in AI? Well, research shows that um in AI overviews, 20% of results include at least one social media platform and that number is even higher in AI mode. Uh 36% of queries will mention at least one social media platform. So the answer to

to this burning question is AI learning from our social presence is a resounding yes. Um AI models are training everything that they can find on social media platforms from the the content that brands put out to the conversations, the discussions, uh the debates and most importantly they use this data to reach a consensus and to kind of determine the truth. So how exactly is AI learning from our social media presence? If in traditional SEO uh you know we we were so accustomed to to keywords, now for AI visibility, they're replaced with brand mentions. So mentions on third-party sites are as important as keywords and backlinks for traditional SEO. It's the new uh vote of confidence. It's the the new kind of seal of approval. Gathering and earning

as many brand mentions as possible is kind of going to assure your your spot in AI visibility. And something that might seem trivial, but the trusted about us pages also play kind of a huge role in whether you show up in AIs and whether the information is is accurate because um AIs really trust LinkedIn and Wikipedia and these about us pages contain structured data that is pre-vetted. So if you if you forget about your LinkedIn page and you have outdated information there, that's what AI is going to pick up and no matter what you say on your own site, you might appear in an AI summary with outdated data. And finally, AI is loves to look at verified uh discussions. We we all figured out that they they like to look um at Reddit comments. They like to follow debate and see what the consensus

is to see what gets up votes and what gets traction. So if we were to look at AI overviews and AI mode, we can see the big dominators here. We have YouTube, Reddit and Kora taking up the most space in these results. And the uh dominators are not too uh too different in CHP. We we know that it relies heavily on Reddit and Wikipedia. So when we're informing our strategy and when we're selecting what kind of platforms we should be focusing on to kind of establish this social presence and make sure that we're feeding AIS the right kind of information. It's worth taking into account AIO's and AI modes and Chip's favorites which are seem to be heavily relying on on Reddit and YouTube. So another burning question, how do AI chatbots decide what to surface? There

are two two ways this uh this kind of surfacing happens. So popular AI chatbots like Chedd and Perplexity actually don't have direct access to to social media, but they do um they do train on anything that's public, anything that's been shared and that was crawled and indexed. So first and foremost, the model um bases the the answers on the internal training data and serves a generalized answer. But if the AI algot has access to um a web search tool, then it's going to perform uh this search and it's going to rely on the traditional search engines ranking algorithm to surface uh a result. It's not entirely clear now what kind of factors uh AI chatbots are taking into account. There's no public data on this, but we are seeing research that's um

indicating what we should be focusing on. And surprise surprise, it it starts with authority. Nothing groundbreaking here if you've been uh around in the SEO world. But yes, LLMs do care a lot about authority and the brand mentions that I was talking about earlier. It's it's more important now than ever to have third party sites mentioning your brand. Whether you're incentivizing that or you're actually doing it yourself, creating profiles and making sure that your company data is is up to date. That's all going to add up and kind of strengthen your entity and ensure that you're serving super accurate data to to AIS. Another super important factor, topical match and depth. Actually, Barry was quoting Kevin Indig a bit earlier, so I'm going to follow in his steps. He said something that really uh stuck with me that this hyper uh selected content

is now becoming as important as it is for users, it's now also essential for AI visibility. So really delving deep into your niche and like owning topics, doing deep dives and really providing that value to to users. This has been said over and over again, but it's still it's still holding true to this day and it's something that LLM really really care about. like they need that expertise and those like super insightful firsthand secrets to surface in in their AI chatbots. And finally, uh we're all going to rejoice in hearing this, but uh traditional SEO still holds. The best practices that we've been using for for years now are still important. And by that I mean alt text and keywords and really optimizing for for the user first and the search engine second. So what part of a post is most

likely to to be read? The foundation is the same both for readers and AI systems first and foremost they they look at the title and the body text because that's what contains the the main story. That's what sets the context and that's what you know that's the the gist of the information. But when the post is audio or video, we did do need to step in and help out with transcripts and captions. And we know YouTube does this automatically, but we can we can enhance captions with keywords and and make sure that we're speaking our users language. And finally, the support is the metadata. So the tags, the alt text, alt text is super important for accessibility and image search. And it puts everything into into context. And it makes sure that we provide a full picture for AI

systems to really understand what the post is about, what other entities it's connected to, and where it would best serve an answer for for the user. A question that I'm sure social media managers are are dying to to learn the answer to is does AI understand visual posts? Yes, it does. Actually, as of this year, models like Chibbt and Gemini, they use optical character recognition and they have super high accuracy in reading text. So, we can all rest assured that our memes are making LLMs uh laugh, not not just our readers. But yes, we don't have to shy away from putting text in our images or taking a more funny approach because AIS can look at the entire context. They kind of connect the dots between the the caption, the description, everything surrounding the text in the image to form a clear picture about what

the post is, what you're trying to put across. And it's all just helping your wider strategy in delivering content that's actually helpful or or funny or more educational. Is engagement a credibility factor? Kind of like backlinks for for SEO. As of now, there's no solid proof to indicate that if a post goes viral, it's going to automatically do well in in AIS. I wouldn't be surprised if up votes likes do have a small small significance there. But in the absence of any like hard proof, I do like to to go straight to the source. And in this case, I went to Google search uh central blog where they say that you don't you don't have to do anything special to to rank in uh in the their AI results. And it it looks like a lot of the fundamentals that we're we're used to are still holding true. like we still need to to focus on delivering value and kind of putting the

the user first and we still need to that evolve at the same time with our users. And now for the measurement and proving the the value that all of this effort we're putting into our social media presence is paying off. At Plannable, we we like to use SE visible. Um I love it because I can see everything at the glance. It's very easy to to have your visibility score at your fingertips to see week over week how your average position is trending. And another super useful feature is the competitive analysis that you can run. You can see how you stack up against your competitors. And you can you can check, you know, what kind of resources they're they're investing in specific assets or content in order to improve their visibility. Okay. So, it's all shaping up to indicate that the SEO team and the social team should work hand in hand.

And historically, they um these two teams have cared about pretty uh different things. But there is common ground. There is we found that on on ourselves at Plannable that both teams do care about authority, brand mentions and entity. And starting from this common ground, we were able to kind of put together a workflow, especially after hearing from other social media managers that we've collaborated with that this kind of like overlap is becoming more and more important not just for for the users, you know, to have that continuity and that kind of consistency across platforms, but also in this new era of AI visibility. So this quarter we've really focused on this close collaboration and we stopped working in in silos. Uh so I'm going to quickly walk you through uh the three-step plan that we implemented at

Plannable. Um not shockingly it started with the planning phase. Basically, we like to to start our quarters with syncs and these are usually made depending on on your team, but we went cross team at this time and we we aligned keyword research with social trends. We established a common keyword database and we mapped out you know the the topics by platforms and and funnels and the result of all this planning was a unified calendar where we could see both the blogs and the social campaigns. Second, the execution velocity. We started working very closely together um the SEO and the social teams and we paired long- form guides with social assets. We started really optimizing social media posts, finding um you know very unobtrusive keywords that we can naturally sprinkle in bios, captions and

and transcripts. And I think most importantly, we didn't try to overdo it. We we listened to our social media manager's uh suggestions and recommendations and we trusted his experience in recommending the best formats. And finally, my favorite step, the the feedback and and measurement loop. We didn't want to do it do like a single meeting and then forget about the whole plan by the end of the quarter. So, we made sure we actually did those uh bi-weekly check-ins. in the team. I'm known that I'm not the biggest fan of meetings. So, this can be check-ins, can be just a five minute check-in or a half an hour chat depending on any new results or anything that's happening in the space. But I've seen that it's super important to share quick wins and have this two-way feedback session because a lot of the times we we find that there's a lot of information in the social team

that we were not aware of trends, comments that gain a lot of traction like very very serious and important topics that might not have search volume, you know, in uh in our in our tools but are gaining a lot of momenting. momentum on social media platforms. So these are the best times to identify these kind of hidden gems and act on them. And of course something that was really helpful was having uh these shared KPIs from the beginning of the planning session. We knew that we were all pulling in the same direction focusing on before traffic and focusing on boosting branded search volume. And yes, this is the three-step plan that has shaped our last few months. And if there's a takeaway I can I can share with you and like a a final actionable step would be to to reach out to your social media manager and ask them to to set up a meeting and start talking about

any any topics that they're seeing across social media. How those are correlating with the keyword research that you're doing and how you can you can start this collaboration. how you can break down the silo and actually you know um ensure that um AI visibility is kind of taken care of and nurtured from the start with both what's happening on the on the block side but also what's happening on uh social media platforms and if there are any questions happy to answer those >> pardon me excellent presentation I truly appreciate the insight for social and how it's applying to AI I and and what we are needing to really adopt and get the teams together. Uh I want to go to Andrea. I want to go to a couple questions from the audience if I could from Clara Bennett. How can social teams create content that's more AI friendly without turning posts into SEO blogs?

>> Yeah. So we actually had this question in our brainstorming session and the best approach here is to kind of stay away from from trying to overoptimize, overgineer, overink and when you when you get your social media manager involved, hopefully they're going to rein in all the optimizations that us SEOs are really eager to do. So, you do you do need to have these two people from the two different teams with kind of different interests find that common ground. Um, and they're each going to, you know, advocate for their their own platforms and their own formats because SEO up until now used to to have the answer and the social team would have the announcement, you know, a blog post and then sharing the news about it on social. I would say that has shifted and now SEO should own the topic and social media should the social media manager

should own and decide on the best format and use their expertise to kind of advocate for the best the best format. >> Excellent. Excellent. Another question from Olivia SEO Carter. If we turn social post into full documentation for AIs to scrape, won't that risk cannibalizing our own website traffic? If the AI gives the whole answer, why would anybody still click through? >> Yeah, that's a valid concern. That's going to happen whether we we help AIS or or not. So why not get ahead of that and actually own the the narrative? Make sure that the data is accurate. Make sure that all the product features are displayed. Everything is is up to date. Nothing is misrepresented and stay ahead of of your competitors because if you're not taking advantage of this, you're just living basically

you're just handing your competitors this opportunity and you might not the might not like the results. The conversation is still happening with or without you, right? >> Exactly. Yeah. >> Excellent. Excellent. Well, Andrea, thank you so much for your presentation, your time