Show impact.
Prove ROI.
Tie outcomes to business goals.
Demonstrate how this drives growth.
Sound familiar? As an SEO, these entirely common and usually well-founded directives were fairly easy to comply with not so long ago. Sources of traffic to your company’s or your client’s website could, for the most part, be grouped into neat little boxes to help you show that your strategic efforts were paying off in the form of clicks and conversions.
And how lovely was it? We could create or optimize a page, it would rank, and then deliver a tidy report showing a direct line to financial growth. Now, with the rise of LLM-driven AI search, the waters of attribution have muddied considerably — at least for now.
Breaks fourth wall: Hi, I’m Shelley, and I refuse to avoid em dashes in my writing to convince you my name isn’t Claude, Gemini or GPT.
What’s Trackable Right Now in AI Search
AI SERPs, assistants and answer engines are evolving into independent distribution channels sometimes outside the bounds of traditional SEO attribution. While some platforms (like Claude and AI Overviews) still leave us guessing, these are the tools you can currently observe in analytics:
ChatGPT
- Starting June 2025, OpenAI added UTM tracking to some links in GPT-4 (web).
- You can also view referral traffic from: chat.openai.com / referral in GA4.
ChatGPT reporting tip: Filter GA4 by session source/medium to isolate ChatGPT users.
Microsoft Copilot
- Trackable via referrals from edgeservices.bing.com / referral and copilot.microsoft.com / referral
- Often shows up when users engage with Copilot via Edge or the Bing sidebar.
Copilot reporting tip: Watch for copilot.microsoft.com in assisted conversions and edgeservices.bing.com as part of multi-touch journeys.
Perplexity.ai
- Includes clickable citations that drive traffic directly to sites.
- Referrals typically appear as perplexity.ai / referral in GA4.
Perplexity reporting tip: Perplexity users tend to exhibit research-heavy, high-intent behavior. Monitor bounce rates and on-site engagement from Perplexity referrals closely to spot potential content gaps, and areas to improve bottom funnel pages.
Google’s AI Mode
Google Search Console (GSC) now includes clicks and impressions from AI Mode in its Performance reports. However, this data is aggregated with standard search metrics and cannot be filtered separately for now. So your clicks aren’t lost — just difficult to trace without a little manual effort.
AI Challenges for SEOs
There are new challenges for SEOs to overcome every year, if not every six months, so let’s jump right into the areas that the rise of AI search has made particularly tricky.
- Attribution Gaps
- Branded vs. Non-branded Behavior in AI environments
- Analytics Limitations
Attribution Gaps and Invisible First Touches
SEOs aren’t new to navigating zero-click searches. Tools like Ahrefs (with its “% of clicks” metric), and Semrush (which highlights SERP feature saturation) helped SEOs and content creators get strategic about the topics they target. This often led to avoiding top-funnel content visibility in favor of longtail clicks to keep reports looking and feeling safe and the same.
In hindsight, this was a tactical workaround rooted in the belief that clicks were the primary outcome to pursue, rather than the mindset shift that was ultimately needed — and would be forced with the rise of AI.
Now, your content might be referenced in an AI Overview without sending a single user to your site. Your brand might be mentioned, your answer might be quoted without ever seeing it reflected in your analytics.
Branded vs. Non-Branded Behavior in AI Environments
Typically, we track brand lift by monitoring increases in branded search over time. But now in AI environments, that brand lift might be happening without a clear entry point.
When users see your brand in an AI answer and search for you later, it’s often tracked as direct traffic or a branded search. At face value, this data in standard analytics won’t connect the dots of your user’s journey.
SEOs must now look for latent signals of recall, like branded search spikes, direct visits, or conversion touchpoints with unclear origin.
Analytics Limitations
Analytics platforms were not built for AI-shaped search behavior. Most AI experiences are non-clickable, no-referral environments. The user gets the answer, hopefully remembers your brand, and returns later in a completely disconnected session.
- Google Analytics 4 doesn’t isolate AI referrers well, with only a few like ChatGPT, Perplexity, and Copilot showing up (and often inconsistently). You can create custom channel groupings to pull in referral traffic, just know that it won’t be the complete picture.
- Google Search Console only recently started including AI Mode data, and doesn’t yet let you filter or attribute them cleanly.
This means SEOs need to lean into:
- Data-Driven Attribution models
- Proxy metrics like branded search
- Qualitative signals like user surveys or customer feedback
How to Adjust Your Attribution
It’s time to redefine what visibility means, adjust your KPIs, and educate your clients or stakeholders about what “impact” looks like in AI-shaped search.
1. Track Your AIO Visibility
Until Google gives us formal reporting, you can rely on tools that monitor AI Overview inclusion ( Authoritas, SISTRIX, SERPwatch, seoClarity, Ahrefs).
- Use these to log how often your content appears in AIO for priority keywords.
- Flag high-visibility, low-click terms where your influence may not register in GA4.
2. Add Qualitative and Proxy Signals
There are steps you can take on your website, as well as cross-channel connections you can make to show that AIO is driving visibility.
- Collect “How did you hear about us?” responses after a user converts.
- Monitor branded search increases after known AIO exposure.
3. Use Data-Driven Attribution in GA4
It’s time to shift from a first-click attribution mindset. Instead of giving all the credit to landing pages, you’ll need to look deeper. GA4’s Data-Driven attribution model uses machine learning to assign partial credit across the user’s entire journey. Without AIO reporting, there are some things we’ll need to infer.
- Go to Advertising > Attribution > Model Comparison
- Set Models to Compare Data-Driven vs. First Click
- Filter by the Conversion Event you want to show attribution for
- Look at Channel Groupings to see if Organic Search’s credit increases after AIO visibility improves.
Example: Let’s say your brand is featured in an AIO response for “best shampoo for hair growth.” Your clicks don’t spike, but branded search queries jump 18% that week and GA4 shows Organic Search playing a greater role in assisted conversions for your hair loss shampoo.
That’s your signal: AIO visibility, followed by brand recall, followed by revenue.
KPIs in the Age of AI
I’m certainly not the first SEO to talk about what KPIs actually matter in the latest evolution of search. (There are roughly 113 new acronyms floating around to back me up on that.) The truth is, AI-driven search demands a shift from measuring just traffic to measuring influence, recall, and cross-channel behavior. Here’s where your focus should be.
Impressions
Impressions are back, baby! Track total and branded search impressions over time to compare against appearance in AI Overviews.
Key Events & Assisted Conversions
When AI tools give answers without clicks, you have to roll your sleeves up and hunt down behavioral echoes of influence elsewhere in the funnel. Watch for spikes in organic and referral traffic credit from Data-Driven attribution after your content appears in AIO.
Placement in AI Overviews
Inclusion in AIO is a win for brand equity, even if it doesn’t result in a click to the featured URL.
- It indicates trust, topical authority, and high content quality
- Track which keywords you appear for and whether branded search or conversions rise afterward
- Log high-AIO, low-traffic keywords for reporting and roadmap decision
Branded Search Growth
AI exposure is the new “I read an article that said…” when they actually saw it on a TikTok. It may not drive a click, but it can certainly drive memory.
- Track branded query volume in GSC
- Compare against periods of known AIO or ChatGPT placement
- Use this as a proxy for user recall, especially when direct attribution fails
Qualitative Signals
When numbers don’t tell the full story, take a page from the marketers of old and listen to the humans. Add a “How did you hear about us?” field to lead forms
- Monitor Reddit, Twitter, Slack groups, and Discord for brand mentions that originate in ChatGPT or Perplexity
- Track brand mentions with Google Alerts
How to set up Google Alerts to track brand mentions:
1. Go to https://www.google.com/alerts
2. Create an alert for strings like:
• "ChatGPT" + "your brand name"
• "Perplexity.ai" + "your brand"
• "AI recommended" + "your product"
• "I asked ChatGPT" + "your keyword"
• "according to ChatGPT" + site:reddit.com
3. Click Show Options and adjust:
• Frequency: “As-it-happens” or “At most once a day”
• Sources: Blogs, Discussions, Web
4. Add multiple variations if your brand is often abbreviated, misspelled, or casually referred to.
Measuring and reporting are only half the battle. Once you know what to track, the next challenge (and the fun part) is optimizing for it without sacrificing conversion potential. Here’s what I think that looks like.
Optimizing Content for AI & Conversions
At the end of the day, clients and shareholders aren’t wringing their hands with worry over impressions and clicks when conversions are happening. The new challenge is crafting content that’s discoverable by LLMs and persuasive enough to drive brand recall, future action, or an immediate conversion.
Here’s how to start.
Ask AI What It Would Recommend You For
If ChatGPT, Claude, or Perplexity wouldn’t recommend your brand for your core offering, that’s not a user problem, that’s a content gap.
Try prompting ChatGPT or Perplexity with queries like:
- “What’s the best [product category] for [use case]?”
- “What are alternatives to [competitor]?”
- “What’s the best brand for [problem your product solves]?”
If you’re not included:
- Update product/category pages to clearly associate your brand with that solution
- Add explicit comparisons, use case coverage, testimonials, and schema
- Make your relevance obvious, repeatable, and machine-readable
Audit Your Perplexity Traffic for Conversion Behavior
Perplexity.ai sends trackable and typically high-intent traffic to your website. You can use that to your advantage.
- Identify which URLs get traffic from perplexity.ai / referral
- Segment those users in GA4 and compare:
- Time on page
- Bounce rate
- Conversion rate
Pages that attract Perplexity users but don’t convert are golden optimization opportunities.
Ask Yourself:
- Is the headline making a promise that the copy doesn’t deliver?
- Is there a strong CTA or product hook?
- Is the page too informational and not actionable enough?
CRO Tip: Heatmap and screen recording tools like Hotjar can help you see where users encounter mental friction in their journey.
Include Semantic Triples in Your Content and Schema Markup
AI doesn’t read like humans, (If it did, it would probably spend all day devouring Terminator fanfiction on AO3.) It parses relationships between entities, concepts, and intent. That means we need to write content that makes clear, structured statements with strong semantic signals. That’s where semantic triples come in.
What are semantic triples?
A semantic triple is a basic structure used in semantic web technology and natural language processing to express a fact. Knowledge graphs, for example, are essentially vast networks of semantic triples mapping relationships between entities.
Semantic triples are constructed of three parts: Subject, Predicate, and Object.
Example: "Vitamin C serum helps reduce dark spots."
That kind of sentence structure:
- Feeds cleanly into how LLMs summarize and cite sources
- Increases your chances of being quoted directly in an AI Overview or assistant response
- Make it easy for the AI to extract your point.
By including this simple, straightforward structure in your content and schema markup, you won’t bury your most helpful statements in overly clever prose (or fanfiction jokes).
AI, Analytics, and the New SEO Playbook
Search experiences don’t end with a click, so it makes sense that success shouldn’t be tied to clicks alone. Success now means maintaining your SEO best practices, taking the steps to show up in AI, and providing content that allows users to convert with confidence when they finally click. Your SEO strategy, your analytics methods, and your content experience all need to evolve together.
After all, it wouldn’t be SEO (and it is still just SEO — fight me) if we didn’t have to pivot at least once a year.
