There is huge opportunity in AI search for local-based businesses
I had the pleasure of being invited to Yext's Brand Visibility Exchange at the iconic Sea Containers location on Thursday morning. This is an opportunity for the Yext team to bring together some of their brilliant clients, all of whom are household names, as well as their partners, to give us updates on the latest in search and in-platform.
The morning's event told a really clear story when it comes to the power of AI and the potential it has to propel businesses forwards. Thankfully, a lot of the key take aways really resonated with what I've been speaking to a lot of clients about lately. Here's the TLDR;
- AI search for local businesses is continuing to evolve, with agents becoming more prominent in models such as AI Mode.
- Leaving no stone unturned with your content; if you're not sharing it, your competitors will, and they will get cited
- Your website is still important, and the majority of the time this is what will be getting cited in LLMs (especially in the likes of Gemini, naturally)
- Google Business Profiles are also still important, and are also cited as often as the website is.
The AI Search momentum continues, and all marketing teams must unite
Speaker: Rebecca Colwell, CMO at Yext
Did you know that the first evidence of marketing goes back to 3,000 BC; in Ancient Egyptian times. Fast forward to now, we're embracing AI after decades of change to the marketing industry.
Rebecca took us through a brief history of marketing, but shared that the fundamentals have stayed much the same, centred around awareness, trust, and control.
The challenge, now, is that we're not just building trust with customers. We're building trust with bots. As a result, we're somewhat starting to lose control of our marketing efforts.
All hope isn't lost though, and one of the key themes of the day really shone through Rebecca's talk which was that things are moving very quickly, and organisations have the opportunity to take advantage of increasingly more data and ways to meet their customers. The key is for all channels to unite and work towards the same shared goal of increasing visibility.
Rebecca spoke in the Q&A about the fact that SEO and PPC should work closely together to review visibility as a shared metric; if Organic is performing well in the likes of AI Overviews, don't push paid there too. Leverage paid to tap into true blind spots.
Citation control categories & proximity research
Speaker: Christian Ward, Chief Data Officer at Yext
One of the starting points of Christian's talk was centred around the fact that any prompt you put into an LLM such as Chat GPT or Gemini, the result you see will be based on your current location if not specified otherwise. This means that if you are working with a client that has a location-varied business, what you are seeing in your results will not be reflective of their different locales.
Importantly, he noted that proximity is incredibly important when it comes to local search results. Within a 1 mile radius of the searcher, a managed [by Yext] business will be +2.71 positions higher. Why? Because businesses appearing in these results are up against each other trying to appear for a visitor then and there. The full report, and the breakdown of the team's ELO scoring can be found here: https://www.yext.com//research/introducing-elo-for-google-rank/study
Christian also discussed the fact that when it comes to citations in different LLMs, there are four categories of 'control' a business will have in order to influence the results;
| Category | Website | Listings; Google Business Profile | Reviews & press; Google, Trustpilot, publications | UGC; Reddit, social profiles |
| Control | Controlled; we control exactly what is included or not included on-site | Some control; we control the structured data, but web users have the opportunity to add photos, FAQs, reviews | Less control; we may have some input into press releases, but no control over reviews | No control; We cannot control what users say about our brand |
| Citation likelihood across LLMs (Average) | 47% | 44% | 5% | 4% |
It was really interesting to see that based on their research, the majority of LLMs are citing sources that are very much within our control; the website and our local listings.
After his talk during the tea break, Christian was very open about the fact that Yext are doubling down on their mantra of openness across the web; they are publishing all of their research at: https://www.yext.com/research
Business success in the future of AI
Speaker: Omar Riaz - Strategic partnerships at Google
Omar set the scene with what we have all come to accept; that users are asking longer (up to 2-3x), and more complex questions in search.
Interestingly, though, he squashed the rumours that younger audiences are moving away from traditional search in Google. He noted that Google search continues to grow, but it is especially growing with younger audiences; Gen Z specifically. We do know, however, that Gen Z are also the social media generation; they are possibly more likely to start their journey within socials, but to still then use Google as they move further down the funnel.
When it comes to Google's AI Mode, Omar shared that this feature is embracing agentic AI. He shared the great example of how users can now see available booking times at restaurants that may have been presented to them in AI Mode as they've searched based on their needs.
It was great to see how much importance was placed on the Google Business Profile (although it would've been strange if it hadn't been, given present company). There were some quick priorities he noted;
- Connect your social profiles to your Google Business Profile
- Aim for 1x weekly Google Post
- Cultivate your reviews (good or bad, you should always aim to reply)
- Photos are important; a high % of customers look at photos before they decide to visit
Much of this we knew, but it was reassuring to hear it straight from the horses mouth.
Is your website still important? Yes.
Speaker: Anna Darova (who also just so happens to have been my very first manager in SEO when I started in the biz!), SEO & AI partner at dentsu
Anna kicked off her talk by giving us some intel on the PPC pilot that dentsu are currently running in partnership with Chat GPT. She shared that based on what they're seeing so far, qualifying queries driving impressions for paid ads are currently low, and the CPCs are high.
She also discussed the fact that paid search costs are rising (CPCs are up ~3-5% YOY), with paid search teams consistently moving more towards AI formats such as AI MAX and PMAX. PPC teams are dealing with highly competitive search, as well as having to explore new formats to reach customers. Sounds pretty similar to the challenges being faced by SEO to me...
I really enjoyed Anna showcasing what a 'perfect page' (for a PDP) would look like for a brand such as Zara, who she noted are not currently being cited particularly well in LLM search. She broke this out into four core pillars:
| LLM Visible | If your content isn't visible to crawlers, then you won't be cited. Simple. This pillar is focused on technical SEO hygiene factors such as minimising JS rendering, improving site performance, leveraging structured data, and adding the layer of auditing server log files to understand the pages that bots are landing on/ |
| LLM Credible | Off-site citations from relevant media publishers, social proof & posting, translation of off-site community feedback (e.g. within Reddit) into on-site content |
| LLM Shoppable | Feeds (both within the likes of Google and also LLMs such as Chat GPT), on-site content, and structured data should all be consistent with the goal of moving closer to an agentic shopping experience within AI. |
Similar to Rebecca, Anna wrapped things up by explaining that it's going to be increasingly important for different marketing channels to become more closely aligned with the shared goal of improving visibility.
Managing your AI agents as you would manage a team
Speaker: Will Warren - Growth Marketing at Yext
Will had a really interesting (and sensible) take on how AI agents should be managed at scale. Essentially, treat them as you would a member of your team.
Put simply; if you give a member of your team a vague instruction, you will likely receive a generic output. If you don't give them any context, they will have to guess. On the other hand, if you tighten the scope, show them what good looks like, and create boundaries, you are far more likely to receive the output you desire.
He detailed three 'titles' of AI agents; the Intern, the Project Manager, and the Manager.
The intern is the one who supports your day to day processes, makes your life easier.
The project manager is the one who has a clearly defined role, that they execute with minimal, or boundaried input. They will know when they need you, and when to take action.
The manager is the one who manages multiple PMs; each of whom know their remit and are executing as planned, with the manager being the one to bring everything together.
Working with AI agents and having agentic workflows isn't something that will happen overnight; Will's suggestion was to start off by rethinking which processes could be evolved. Plan what the role would be, as you would for a hire; define what you will need them for, when they need to be involved, what personality traits they would have, and so on. That way, you're already beginning to build a picture of what their output should be, and what their boundaries are.
Anna Darova talked about why your website is still so important
In summary
One of my favourite things about the morning was the fact that all speakers and attendees were telling the same story and ultimately singing from the same hymnsheet. We all truly believe that the opportunities for our clients and partners is great, and that we're all on that same journey together.
As search expands, we need to work closely with wider marketing channels in order to be present across different platforms. Our customers are going to get increasingly more used to having all information at their fingertips, and so businesses need to review their strategies and ensure they are understood by bots and customers alike.