Tim Soulo: Where Marketing's Headed And How Ahrefs is Evolving | Ahrefs Evolve 2025

Ahrefs Tutorials| 00:29:56|Mar 25, 2026
Chapters8
Tim introduces himself, clarifies his role as CMO, and sets the stage for discussing marketing leadership at Hrefs.

Tim Soulo explains Ahrefs’ evolution, AI-driven strategy, and new tools like Brand Radar and Hrefs Labs to help brands gain AI visibility and discoverability across search and social channels.

Summary

Tim Soulo, Ahrefs’ CMO, shares the origin story of Ahrefs (formerly hrefs.com) and outlines how marketing at the company has evolved over a decade. He emphasizes that the product’s daily data updates outpaced competitors early on and sets the stage for Ahrefs’ AI-focused future. Soulo dives into how AI is reshaping search, clarifying that AI overviews, not pure keyword cannibalization, are changing traffic dynamics, with 21% of Americans acting as active AI tool users but Google still driving the majority of traffic. He introduces Brand Radar and Brand Raider for AI visibility, citations, and entity tracking to monitor how brands appear in AI chatbots and across prompts, with a growing database of prompts (over 10 million for CH GPD) and upcoming features like custom prompts. The talk covers practical AI integration in content creation, including an AI content helper that scores topic coverage and a brand kit to preserve tone of voice. Soulo also previews Ahrefs Labs as an internal R&D hub to automate marketing processes, such as an automated monthly blog report, and MCP servers that connect to AI chatbots to pull Ahrefs’ data. He ends with a strong message: AI-assisted SEO is not a threat but a broader marketing paradigm; Ahrefs is expanding into social, video, and AI-enabled discovery to keep brands visible “everywhere.”

Key Takeaways

  • Brand Radar provides AI visibility metrics, citation tracking, and entity-based analysis to compare brand presence across AI chatbots and prompts.
  • Ahrefs’ Brand Raider uses a large prompts database (10M+ for ChatGPT and growing) to quickly assess brand visibility in AI systems without manual prompt generation.
  • AI content helper analyzes drafts to reveal topic coverage gaps with color-coded visualization and offers an AI-assisted workflow that enhances, not replaces, human content creation.
  • Brand and product updates include AI overviews in site explorer, MCP servers for AI-assisted data, and Ahrefs Labs to prototype marketing automation (e.g., monthly blog reporting).
  • Ahrefs believes SEO is expanding into a broader marketing ecosystem—discovery across AI, social, YouTube, and Reddit is increasingly important, not just traditional search.
  • Ahrefs’ free web analytics tool is central to their strategy, boasting 100k+ installs and ongoing feature updates (custom events, funnels, 404 reporting, etc.).
  • Content strategy with AI prioritizes meaningful storytelling, data-driven experiments, and fresh discoveries over simply generating more content.

Who Is This For?

Marketers, SEO professionals, and product teams who want to understand how AI shifts search and content strategy, and how to use Ahrefs’ new tools to stay discoverable across AI and social channels.

Notable Quotes

"“Href is HTML for a backlink.”"
Tim recounts the quirky origin of hrefs.com and how branding emerged from Mitro’s engineering-driven pivot.
"“AI overviews are here to stay.”"
Tim highlights the impact of AI-generated overviews on search traffic and how Google still drives most traffic.
"“Content creation isn’t the goal of AI; improving content with AI is.”"
He reframes AI in content as a tool for better storytelling and data-driven experiments, not just automation.
"“Brand Raider works super well when you have a bigger brand.”"
Tim explains the scope and limitations of their AI tracking tools for brands of different sizes.
"“Make your business discoverable in search AI and beyond.”"
Closing tagline reflecting Ahrefs’ expanded vision across traditional search and AI-enabled discovery.

Questions This Video Answers

  • How is AI changing Ahrefs’ approach to SEO and content strategy in 2025?
  • What is Brand Radar and Brand Raider, and how do they help with AI visibility in chatbots?
  • What are MCP servers and how do they integrate with AI chatbots for Ahrefs data?
  • How can I use Ahrefs’ AI content helper to improve article quality without overusing AI?
  • What is Hrefs Labs and what marketing automations are they prototyping for marketers?
Ahrefs origin storyBrand RadarBrand RaiderAI visibilityAI overviewsBrand citationsMCP serversAhrefs LabsAI content helperfree web analytics
Full Transcript
Apparently, I had to disappear to appear again. Uh yeah, my name is Tim Solo. I'm chief marketing officer at Hrefs, not the founder. Some people think I'm the founder. I'm not. Uh I joined the company 10 years ago back in 2015. There there were just 15 people in the company total. I was the only marketer. I quickly switched uh snatched the title of CMO so that no one else could get it later. Uh 10 years later, I'm still at HFS. Had zero career progress. But yeah, it's not about me. We're talking about where marketing is headed and how HFS is going to help you navigate it. But first, I want to share a little bit of backstory of how Hrefs came to be because I actually think it's quite funny. So, this is our CEO and founder Mitro. And back in 2010 he was working on some programmatic SEO projects before programmatic SEO was a buzzword in our industry. And back in the days those file sharing sites like media fire, zippy share, mega upload, they were quite popular. They were generating a lot of traffic. And Mitri had this idea let me do kind of file hosting and search for PDF files. But for that he needed an index of all PDF files that existed on the internet. Where do you get it? There was a company there still is a company called Majestic uh that have an index of the web. So Ditri went to them. He said how much is your API? They gave him a quote. He was surprised. And being Dimmitri in Ukraine being like super talented and super smart engineer he thought this is how much people are paying to get this data. We're pivoting now we're building our own majestic and this is basically how HFS came to be. Uh Mitro and Eager our CTO they started working together. They built an amazing product and Dimmitri thought how should I call this company that I'm starting. Well, links.com was taken. So, Dimmitri being an engineer chose hrefs.com. As you may know, href is uh HTML for a backlink. As you can tell, as as Ditri joking himself, he didn't have a marketing person back then to help him with some other brand names. So, HF.com is Yeah, it's uh Href in HTML. 2011 website goes live. Uh and the thing is while Majestic which Dimmitri kind of copied the idea from was updating their database every month, HFS was updated every day. People immediately see the difference and the product started growing like it didn't take a lot of marketing for people to started using to starting to use us and recommend us to their friends. But I remember that when I connected with Mitri back in 2015, he sent me this Slack message of his kind of vision of the market and he said that Hrefs in his opinion had tech five out five out of five, marketing one out of five. So my job was to bring our marketing to five and u I don't want to put uh to pat myself on the back too much but I think it has worked pretty great. great product, great marketing, and we took off. Uh, since our company is headquartered in Singapore, we're actually moving to US, but that's another story. We were consistently hitting the list of Singapore fastest growing companies. So, the past 10 years were quite chill until, as you know, AI hit our industry. And AI is changing two things. It is changing how we search and it's changing how we work. So I want to discuss this with you today starting from how we search. I recently shared this screenshot on LinkedIn uh like a crocodile screenshot of our Google search console data. It got a lot of traction because a lot of people were seeing the same thing in their Google search console. Uh and Patrick has already talked a lot about this so I won't dwell on this too much. Basically a lot of people imply that it is chat GPT that is beating beating Google. This is why our clicks are dropping. Our traffic is dropping. That is not the case. Again the study that everyone is referencing by Renfishkin great study. Google is still going strong. Chad GPD is a lot smaller than Google. Another cool study that Renfishkin did with clickstream data is that their study that the adoption of AI tools uh among Americans specifically is growing. As of today, 21% of Americans are active users. I think to qualify as an active user, you had to do like 10 queries on Chad GPT per day or something like that. Again, 21% are active users, but it is not at Google expense. Still 86% of Americans are heavily using Google. So it's not Chad GPT who stole our clicks. It's AIO reviews. Google itself is stoalling traffic from us. And uh we did a great study at Trevs. Uh Ryan Law did it together with our data science team. We crunched the numbers and we identified that the presence of AI overviews is dropping clicks by as much as 34%. That's quite a lot. But the thing is Google themselves they're quite happy about it. They they're saying that AI overviews in search increase search users and increase user satisfaction. So what we can get from this is that AI overviews are here to stay. Quick plug because we're at HF's conference and I'm talking about HFS. Prepare yourself for a lot of HF screenshots. uh we have AIO reviews filters uh in both our keywords explorer and site explorer tools because there are still a lot of keywords where Google is not displaying any AIO reviews and those keywords still have a lot of opportunity to get traffic but overall we've been seeing kind of an evolution of zeroclick searches in the past decade so it started from featured snippets and other SER features that Google was displaying to give people something before they click Then we had Chad GPT and Plexity. Then Google answered with their own AI overviews. Now they're rolling out AI mode. And who knows what would happen in the future. But one thing is clear, there would be less search traffic going forward because it seems that those systems are incentivized to keep people within themselves and not give traffic away anywhere. But at the same time, they're going to be more searches because those systems allow people to do more things. Here's the thing though, AI chatbots because they want to keep users within their kind of world garden, they won't send you much traffic. Uh again, Patrick referenced this fun website that we actually vibe coded, chedg.com. is based on I think at this point 70,000 websites which are connected to our free web analytics product and we can see that yeah the traffic that chat GPD is sending to websites is slowly growing but as of right now it accounts for just 0.24% 24% of all referral traffic that websites get. Compare that to Google, which drives like 41.35. So, it's nowhere. Yeah, AI chatbots won't send you much traffic, but they can recommend your your brand and your products. So, here at Hrefs, we have selfattribution. Every time someone signs up for a free account or a paid account, we ask them, "Where did you hear about Hrefs?" And so far we have over 30,000 people that mentioned chat GPT when they were answering this question. And it's not just us who are noticing this. So the other day I saw a tweet by Nathan Bchess. I'm butchering his last name probably, but bear with me. He said, "Increasingly we have signups saying they found us through chat GPT. Is anyone building HFS for LLMs? I couldn't resist answering by HFS is building HFS for LLMs. So, enter brand raider. This is HFS for LLMs. Basically, it works the same way as site explorer. You put in a website uh but in this case you put in a brand and we instantly show you the visibility of this brand in different AI chatbots and you can compare a bunch of brands together. And you can see how their relative uh AI visibility compares. There's actually a fun game that Sam was playing with a bunch of attendees of this conference where they had to pick the popularity of different things as seen is as seen is in AI overviews. So that would probably go on our Tik Tok and Instagram account later. If Sam asks you participate to participate in this thing, please participate. It is quite fun. So yeah, you can compare you can compare visibility of different brands instantly. So you don't need to come up with the list of prompts that you want to uh to get data for. We have a massive database of prompts. I think for CH GPD it's over 10 million now and we're growing it over time. So we give you a good snapshot. You don't have to wait. You can input any brand, any word, anything. We would give you a snapshot of how popular it is uh in AI search. You can also work with entities rather than brand because some brands they can be misspelled or there are multiple ways you can write them out. For example, crazy egg uh without space and with space you get different results but you can track it as an entity. And finally, the most important thing in optimizing for AI search is of course citations because of course there's training data where AI chatbots are pulling information from but that training data has cut off time. So you cannot really influence that because that was in the past. What you can influence is the websites where it is pulling fresh information from. It seems that in our industry we are now calling this citations. And for any brand that you're researching in brand radar, you can get a list of citations where uh AI chatbots are pulling information from. And your job right now is kind of to do online reputation management. You have to work with these citations and make sure that information about your brand at these websites at these pages is accurate, upto-date, uh ample and all of that. So yeah, like I said, brand radar, it works like a chef site explorer. So you don't need to give it a list of prompts you want to track. It already has a massive index of prompts. So for AI overviews, it's the biggest. For ChatGpt and others, it's around 10 million. Uh but we're expanding it all the time. Custom prompts are coming because I'm not saying that there are not useful. They are very useful because Brand Raider works super well when you have a bigger brand. For smaller brands, of course, our database is not yet there. Even though we're expanding it, it's not yet there to give you a lot of information about smaller brands. So, in about a month or two, maybe, don't take my word for that, please. Uh we're going to release custom prompts and you'll be able to track anything you want in different chatbots. Overall, we're seeing that all marketing teams are obsessed with their AI visibility, giving the trajectory of its of it all. Uh, a lot of people are thinking that AI search is our future. So, you want to start working on your AI visibility right now. And this is why all those metrics, how you show up, what's your AI visibility in different AI systems, AI chatbots is now part of our site explorer because we think that it is very important to not just know your traffic, keywords, backlinks, but also your AI visibility. But brand radar, it's not just AI visibility tool. Uh again if you listen to all other speakers SEO is starting to become more like marketing you need to look into more different areas. So, brand radar right now it has a AI visibility index, it has search demand index, it has web visibility index, and we're also adding YouTube and Reddit soon because for your brand, we want to give you kind of a 360 view of what's happening because all those things are interconnected. Like Patrick said, it is all just marketing. You have to do a little bit of everything for AI systems to know about you and to pick you up. So that was how we search. Now how does AI change our day-to-day work? What we actually do at work. So the first thing I want to make clear is that LLMs because AI there's different AI. There's predictive AI and that's like a whole different thing. But LLMs are essentially text generation machines. So when marketers hear text generation machines, we're like, let's generate some text. Is it a good approach? Let's see. So, uh, if you know, I started HF's podcast about a year ago and I've been interviewing some interesting people. I actually to be honest with you I do this for myself because it's a good excuse to have someone I'm interested in and ask them all sorts of challenging questions where like if I just tell them can I hop on a call with you and grill you with my questions and you get nothing in return they would probably think I'm crazy but then I'm saying oh like it's for publicity they suddenly open to do it. So I had Zach Harris he used to work at copy.ai AI kind of the epitome of creating content with AI and he positioned himself as a huge advocate of creating AI generated content and I really wanted to dig into this. I wanted to understand how good of a content piece you can generate with AI and right during this podcast interview I came up with an idea of a challenge. So, the challenge was that we're taking the transcript of the podcast that we recorded with Zach and he is going to use AI without any human intervention to turn the transcript of our podcast with him into two articles. And I would use human neurons to do essentially the same thing. So, I would hire a bunch of freelancers, not our in-house team. I didn't want to distract our amazing people to write articles based on our podcast interview. But that was basically the bet. We wanted to see who can produce better content. So then uh I asked Michelle, our podcast producer, to share those articles within our own team and have people vote which articles they thought were AI, which articles they thought were written written by human beings. And uh as you can see for one of the articles people clearly saw that one of them is AI and they were right. On another article it was quite mixed. But the thing is that was the wrong question to ask in the first place because all four articles were Like uh I'm not a native speaker. I'm Ukrainian, so I'm not sure if I can say the word from stage, but I already did it twice, so it's okay. So, yeah, that's that's the wrong question to ask. It's it's not about AI or about the quality of content. Like, what is quality of content? Creating content is the wrong goal to start with. And our goal from this podcast interview, our bet was like, can we create content out of this podcast interview? That's the wrong goal. What is the right goal? It's to teach what you know. It's to share fresh ideas. It's to break big news, tell great stories, run bold experiments, study real data, make new discoveries. This is how you create amazing content. Your goal shouldn't be to ask someone, can we create more content? No. Do we have any cool stories we can tell people? Do we have any interesting data we can share with people? Do we have any experiments we can run? This is what we should be doing. So, when you have a goal to create content and you add AI, I'm not going to say this word again, but you see what happens. But when you have the right goal and you add AI, then you basically have magic because your content, you can improve your content. AI can help you push your content to the next level. Speaking of which, this is why we have AI content helper with HFS. It doesn't let you generate content. So you cannot go and say generate me content about this and you'll have an article. This is not this is not in our philosophy. This is not what we believe in as a company, but it helps you improve your content. How? It scores your content for a bunch of topics. So, it breaks your piece of content into individual topics that are being covered and it gives you score of the coverage of each each individual topic and helps you understand the gaps. And what's quite convenient is that it actually has color coding because like Patrick said, you don't cover different subtopics of your piece of content in sequence. You can actually see like which part of your content are covering this topic and which parts of your content are covering that topic. It's very visual. It's very clear to see where you have potential gaps and what you can improve. Other than that, of course, we have kind of AI assistant within our AI content helper. This is just UI because yeah, you can you can do stuff in chat GPT. You can like put your your article your draft into chat GPT and ask it to improve things. But also it's very convenient when you have UI made by people who know their stuff and who are actually using this ourselves. Recently we added brand kits uh which is a very cool feature which helps you to retain your tone of voice and how it works you actually give it a bunch of articles. You give it uh your website name and it generates kind of like a system prompt uh with a lot of information what it knows about you and you can then go and edit it if you disagree with something. it does and then you can fine-tune it. If you think when it it helps you generate uh some paragraphs, if it is like a bit a little off, you can go back to the system prompt and change it. And the most important thing is that our own content team is using AI content helper. I specifically messaged Ryan and I said like actually this was like more than half a year ago. I think we're using it even more as we as we add more features. But he said that it was super useful for help outline new posts because you can just uh study your competitors and it will give you a quick outline of what you should cover in your article which can serve as a basis and then it is very cool for updating existing content because if your article is outdated, it can show you the gaps and you know what you should update. Yeah. The thing is using AI to create content is no longer optional. Right now it's table stakes. We actually studied how much content is created or assisted with AI. So what we did this was in uh April 2025. We took about a billion pages that our web crawler like so freshly created and we ran these billion pages through our AI content detector to see how many of them were like pure AI, how many of them were pure human content and where was the mix and the mix can go anywhere from like heavy AI usage to very little AI usage. Uh and yeah, what's interesting there's not that much pure AI content. Most of it is kind of AI assistant. So the question is how much AI is too much? How far can you scale it? Which content types are more forgiving to AI intervention? And where is AI detrimental? This is the main reason why we have AI content detector tool which can scan any page which actually already scanned a lot of pages of a lot of websites and can show you not just that the page is using AI but again it can color code where exactly AI was used and which AI systems were used because there's like glam GPT Gemini blah blah blah uh yeah see where exactly AI was used in your content and Um, I was talking I think it was we had dinner with Joe the other day and he said that it is so awesome that you can put any website into HF site explorer. Go to top pages report and see how much AI was used in creating those pages and how much traffic they get because this helps you answers those questions like if I go overboard with AI would it tank me or not? and you can analyze your competitors and you can see how much AI they're using and you can make your own decisions. What else HFS is building for the new AI era in marketing? Well, we talked a lot about MCP servers and a bunch of speakers already mentioned that we have MCP server for HFS. This is the thing you can connect to uh AI chatbots like chat GPT Gemini and you can ask it to do stuff with HF's data and there's a lot of uh connectors that you can put pull data from HFS and Glenn Alop our recent awesome hire. Can I get a round of applause for that? Do you know Glenn Alup? Yeah, I'm kind of I'm kind of pretty proud of getting al Glenn al to come work at HFS. So yeah, literally yesterday, it was Monday, right? He published an article with 15 HF MCP use cases. So if you're new to this, this article gives you quite a nice overview uh of how you can use HF's data in your AI chatbots. The next thing which is quite interesting, we haven't released it yet. Uh internally, we're calling it HFS Labs. Basically what it is uh since I joined the traps for quite a few years I was asking our founder Ditri can you give us marketers our own developer so that we wouldn't need to ask the product team to build something for us so that we would have like our own little kind of research and development department R&D department for marketers and now with the advent of wipe coding where kind of prototyping something becomes super past Dimmitri himself said guys this is what you should be doing now I'm giving you even like a couple developers and you're going to work on something called Hrefs Labs whatever processes you have in your marketing department that are repeatable processes that you think can be automated with Href's data with AI you have to build them within HF's labs we will release it to our customers we'll see how it goes and then if it picks up speed we'll we'll continue developing it. So the first project that we're working on uh under HF's labs is we're automating our own monthly blog report. So because we have fairly sizable content team, I think it's five people who are contributing to our blog if we're not counting uh people who are also work on the product and Ryan every month he has to create a monthly report with everyone's performance. And it's not just for me kind of like as a CMO like Ryan show me the report of the blog. It's for everyone to understand how they're doing, what's their output, how is their content helping the company to grow and all that. And Ryan of course being quite a smart guy, he also took a Python course. He created some scripts to automate his reporting. But then we thought not every content lead at every company is going to do this. So we need to build this for people. And yeah, this is just the first project. Uh we have a lot of ideas in our marketing department. Now that we have Glenn, we have like overwhelming amount of ideas of what we can do and we'll be shipping them in HF's labs. So, this is something to uh watch out for, to look out for. Sorry for my English. One last thing, you get the vibe of Steve Jobs, right? Free web analytics. I mentioned it quite a few times. This is actually a real screenshot of the uh HFS evolve website. This is how we were promoting our event and generating traffic. But yeah, uh we are a big users of our own products. I'm personally loving our free web analytics. Uh it's free for websites with verified ownership. It's privacy friendly, zero complexity, and since the time we launched it, I think in the beginning of the year, I I actually wrote uh a LinkedIn post listing some of the most important features that we added, things like custom events, funnels, GSC integration, cloned pages report, possible 404 report, which Patrick talked about recently, filter presets, and that's not even all of them. So, we're shipping updates to this product like crazy, and this is free. We're using it ourselves. So, if you need a simple uh analytics solution, make sure to sign up. Yeah. So far, we have over 100,000 installs in about 10 months since we launched it. Two last thing. I told you I'm I'm Ukrainian. My English is a little bit off, but I promise there would be no three last thing. That's that's it. Uh social media manager. We launched the social media manager. Uh it's pretty rudimentary right now. So it has like just the basic features. You can schedule posts to go out uh at a few different social networks. The next thing is coming is that we're start we're going to start giving you some interesting analytics about the performance of your posts about like who on your team posted that what's the output. Again, this is something we need internally because as you might have noticed, I'm quite active on LinkedIn, Twitter, we have other team members, Ryan, Patrick, we have our branded accounts, HFS on X, HFS on LinkedIn. So, we're outputting quite a lot of content and we need this tool for our own team and we want to build some reporting for ourselves and some tools that would be helpful for us. This is how a lot of products in HFS come to be because our own marketing department needs this. But when I announced that we're releasing social media manager uh on Twitter, someone by the name Dan said, "Interesting pivot. Is the SEO market contracting?" Uh is Dan in the room by any chance? Dan? No. Dan? No, I don't think SEO market is contracting. I think SEO market is expanding because like Patrick said and like many other PE speakers today and tomorrow are going to tell you SEO is becoming more like marketing. So, at first we just cared about Google, maybe Bing, maybe Duck Duck Go. Then we thought, how do we get our content on YouTube? How do we get on Google Maps? How do we show up on Reddit? Now we care about social networks because there's a lot of attention. Marketers are hunting for attention and there's a lot of attention on social networks. So, can people find you on social networks? And now we're also adding AI chatbots. Like I said, Google isn't really losing market share, but AI chatbots is yet one more thing where you can show up. So, the SEO market is actually expanding, which is good for everyone of us in this room. Thank you. And HFS is not pivoting anywhere. We're just building tools to make your business discoverable everywhere. And so you might have noticed that on our homepage we changed our tagline. Now it says make your business discoverable in search AI and beyond. This is it from me. I hope you enjoyed it. Uh I also need to make a selfie with you guys. Can can you light up the room a little bit? Okay. Can you raise your hands? Maybe like you enjoyed my talk. Awesome. Thank you guys.

Get daily recaps from
Ahrefs Tutorials

AI-powered summaries delivered to your inbox. Save hours every week while staying fully informed.