We spend a lot of time talking about AI search visibility. We track brand mentions, monitor AI citations, and look for ways to get our clients’ content surfaced in ChatGPT, Claude, and Google’s AI experiences.
But there’s another question worth asking.
Which AI bots are actually visiting our websites, and what are they looking at?
That’s what caught my attention when I started testing OtterlyAI’s new Agent Analytics feature. It offers a look at the AI agents crawling your website, the pages they’re visiting, and the different engines behind that activity.
I connected my personal WordPress site to see what the data would reveal. Here’s what I found, how the feature works, and what I think SEO professionals can do with this information.
What Is OtterlyAI Agent Analytics?
OtterlyAI is an AI search visibility platform with research tools designed to help marketers understand how brands and content appear across AI search experiences.
Agent Analytics adds another layer to that picture by showing website owners which AI agents are accessing their sites.
Instead of looking only at rankings, traffic, or AI-generated answers, you can examine the crawling activity happening behind the scenes.
For SEO professionals, that raises some interesting questions:
- Which AI agents are visiting my website?
- What pages are they crawling?
- Are different AI engines accessing different parts of my site?
- How can I use this information to inform my SEO and content strategy?
The feature doesn’t answer every one of those strategic questions on its own. But it provides another source of data to investigate.
How to Set Up OtterlyAI Agent Analytics
The first step is connecting your website to OtterlyAI. The platform offers several data source options, including Cloudflare, Webhook, WordPress, and file upload. You can also request a new integration.
I chose WordPress because it was the most straightforward option for my personal website.
Connecting a WordPress Website
To connect a WordPress site, you install the OtterlyAI plugin and follow the setup process.
Once your website is connected, you can navigate to Brand Reports and open Agent Analytics to explore the activity.
For my test, I used my personal website as a sandbox. This gave me a way to explore the feature without having to request additional access or credentials from a client.
If you’re an SEO consultant or work with multiple clients, testing new tools on a website you control can be a useful way to understand the data before introducing it into a client workflow.
What Can You See in Agent Analytics?
Once your website is connected, Agent Analytics provides an overview of the AI agents visiting your site.
One of the more interesting aspects is that these agents don’t all serve the same purpose.
1. Crawlers and Indexing Agents
Some bots crawl websites to discover and process content.
This is where technical SEO fundamentals still matter. Your website needs to be accessible, your important pages need to be discoverable, and your internal linking and sitemap should make it easier for crawlers to navigate your content.
During my walkthrough, I noticed that the sitemap was one of the pages receiving crawler attention.
That makes sense as a technical SEO consideration, although a bot visiting a sitemap doesn’t automatically mean every URL listed in it will be crawled or indexed.
A clean XML sitemap, an accurate robots.txt file, and a logical site structure are still worth reviewing as part of your technical SEO process.
2. AI Training Data Scrapers
Another category includes bots associated with collecting content for AI model training.
This activity is different from a bot accessing content to support search or answer-generation experiences.
It’s important to distinguish between the two. A crawler visiting your website doesn’t necessarily mean your content will be cited in an AI-generated answer, and training-related activity isn’t the same as a referral visit from an AI platform.
The distinction matters when interpreting crawler data and deciding what it actually tells you about your website’s visibility.
3. User-Triggered Agents
User-triggered agents are particularly interesting because they can be associated with activity initiated by a user, depending on the agent and how it operates.
For example, an AI tool may retrieve a page when someone asks a question that requires current or specific information from the web.
This is a useful area to investigate because it connects crawler activity to the broader question of how AI tools access online information.
However, crawler logs alone don’t tell you exactly what a user asked, whether your content appeared in an answer, or whether that interaction resulted in a website visit.
Those are different signals, and they need to be analyzed separately.
Which AI Engines Are Crawling Your Website?
Another feature I found interesting was the ability to explore activity across different engines and agents.
When you open the analytics, you can drill into individual pages and examine the agents associated with their activity. You can also look across the available engines to get a broader view of the bots accessing your website.
The number of different agents was surprising.
It was a reminder that website discovery is no longer limited to the traditional search engines most SEOs have spent years monitoring.
However, I noticed something unexpected during my test: Google appeared to account for less activity in this particular view than I would have expected based on other analytics data.
That raised an important question about how the different tools measure activity.
Google Analytics, search performance reports, and server-side crawler analytics measure different things. They aren’t interchangeable, and their numbers won’t necessarily line up.
Before drawing conclusions, I’d want to understand precisely which agents are included, how activity is classified, and what time period each report covers.
The takeaway isn’t that one platform is necessarily crawling more or less than another. It’s that each data source provides a different view of the search ecosystem.
What Does AI Crawler Data Mean for SEO?
This is where I think the conversation gets more interesting.
Knowing which bots are visiting your website is useful information. But data collection isn’t the same as strategy.
The next step is figuring out what the information can tell us, what it can’t tell us, and how it fits into the larger picture of organic visibility.
Here are three ways I’d approach the analysis.
1. Investigate Which Pages Attract Crawler Activity
Start by looking at the pages receiving attention from different agents.
Are they primarily visiting your homepage, product pages, blog posts, documentation, or other resources?
Then compare that activity with your existing SEO priorities.
For example, if an important educational page isn’t receiving the crawler activity you expected, you could investigate its accessibility, internal links, sitemap inclusion, and technical status.
That doesn’t mean low crawler activity automatically indicates a problem. It gives you a reason to investigate.
Likewise, a frequently crawled page isn’t necessarily performing well in AI search. You’ll need other data to understand whether that page is being surfaced, cited, or generating traffic.
2. Compare Crawler Activity With AI Visibility
Crawler analytics becomes more useful when you evaluate it alongside other signals.
For example, you could compare the pages receiving crawler activity with the pages appearing in AI-generated answers or being cited by AI search tools.
This is where AI visibility research can help inform the analysis. The goal is to understand where your brand appears in AI search, which sources support those answers, and where opportunities may exist to strengthen your visibility.
The two data sources answer different questions:
- Crawler analytics: Which bots are accessing your website?
- AI visibility research: Where and how does your brand or content appear in AI-generated answers?
- Website analytics: Are people arriving from AI platforms, and what happens after they visit?
Looking at these signals together can provide a more useful picture than relying on any one metric.
It’s still important to distinguish correlation from causation. A page being crawled doesn’t prove that the crawling caused an increase in citations, rankings, or traffic.
3. Use the Findings to Inform Content Strategy
Suppose you discover that certain educational pages receive activity from several AI agents.
That could be a starting point for further research.
What questions does that content answer? Are there related topics your website doesn’t cover? Does the page clearly explain the subject, demonstrate expertise, and provide useful information that other resources may be missing?
You could then compare those findings with traditional keyword research, competitor analysis, and AI visibility data.
This is where an AI SEO strategy can bring the different pieces together: technical accessibility, content quality, search demand, and visibility across emerging search experiences.
The goal isn’t to create content simply because a bot visited a page. It’s to use the data as one input into a broader strategy that supports your business goals.
What Agent Analytics Doesn’t Tell You
As with any new SEO tool, it’s worth being clear about the limitations.
Crawler activity is a useful technical signal, but it doesn’t directly measure AI visibility or business impact.
A bot visiting your website doesn’t necessarily mean:
- Your content was cited in an AI-generated answer.
- Your brand was recommended to a potential customer.
- A user clicked through to your website.
- Your page influenced a model’s response.
- Your content generated a conversion or revenue.
Those outcomes require different forms of measurement.
For example, you might use AI visibility tracking to monitor brand mentions and citations, then use Google Analytics 4 to investigate referral traffic and conversions from identifiable AI sources.
Even then, attribution can be incomplete. Not every AI-assisted interaction results in a trackable referral, and not every visit can be attributed to a specific AI answer.
This is why I wouldn’t treat crawler counts as a standalone KPI for AI SEO success.
Instead, I’d use them as a diagnostic and research signal alongside other performance data.
Is OtterlyAI Agent Analytics Worth Exploring?
My initial impression is that Agent Analytics is an interesting addition to OtterlyAI’s research capabilities.
The setup was straightforward using WordPress, and I appreciated being able to explore the different agents accessing my site.
The most valuable part, for me, is the additional visibility into website crawling activity. It gives SEO professionals another data source to investigate as AI-powered search continues to evolve.
The next challenge is turning that information into practical recommendations.
I’d like to see how the feature develops and how its data could support more actionable insights for SEO strategists. Understanding which bots are visiting is a useful starting point. Connecting that activity to content opportunities and measurable business outcomes is the bigger opportunity.
If you’re already using OtterlyAI, I’d recommend exploring the feature on a website you control and seeing what you learn.
Just remember: Interesting data is the beginning of the research, not the end of it.
Final Thoughts
AI search is creating new questions for SEO professionals, from how content is discovered to which sources are cited in AI-generated answers.
Tools like OtterlyAI Agent Analytics give us another way to investigate what’s happening behind the scenes.
But the real value comes from combining that information with technical SEO, traditional search research, AI visibility monitoring, and a clear understanding of business goals.
We don’t need another dashboard full of numbers without a plan for using them.
We need to know which signals matter, what they tell us, and what decisions they can help us make.
That’s where SEO strategy comes in.
This review reflects my initial exploration of OtterlyAI Agent Analytics. Features and capabilities may change as the product develops.
