AI search is getting harder to predict
Answer engines like ChatGPT change which businesses they mention from day to day. Here's how to tell if it's the engine or your fault.
AI answer engines are becoming less stable. ChatGPT, Google Gemini, and Perplexity now mention and cite different businesses from one search to the next, even when someone asks the same question twice. This matters because more marketers are trying to get their content into these AI answers instead of just ranking in Google. The challenge: when your traffic from AI engines drops, figuring out whether the engine shifted or your strategy failed requires comparing your numbers to industry-wide data.
Key facts
- In the first quarter of 2026, 58 percent of marketers were working to get their content into AI answer engines.
- When researchers asked ChatGPT and Google's AI system the same question repeatedly in late 2025, the engines returned the same list of brands less than once in a hundred tries.
- AI search volatility refers to the unpredictable swings in which brands answer engines mention and cite, separate from regular Google ranking changes.
- The randomness is built into how these language models work; they are designed to produce varied answers rather than identical ones.
Why this matters for your business
If you have been trying to get your business mentioned in ChatGPT or Gemini answers, you are now chasing a moving target. A tool that worked last week might not work this week. More important: you cannot track your performance the way you can with Google search results. Google gives you a rank you can monitor. With answer engines, you can track whether you are mentioned, how often you are cited, what type of citation you get, and how much traffic they send. But all four of those numbers can shift without warning.
How to tell if the problem is the engine or your strategy
When your traffic from AI answer engines drops, compare what happened to your business against what happened across your entire industry during the same period. If every business in your field saw a traffic drop at the same time, the engine changed something. If only you dropped while competitors held steady, the problem is likely your content, your website structure, or how you are tracking the traffic. Tools now exist to show you industry-wide trends so you can make this comparison. Without that context, you are just guessing.
What to check in your own content
Answer engines favor certain types of content over others. Blogs, comparison pages, guides, and documentation get cited at different rates depending on what the engine currently treats as authoritative. If you notice that a particular content format is growing in AI answers week after week, that is a signal the engine is favoring it more. If a format is shrinking, you may have less reason to keep investing in it. Watch for these shifts and adjust where you spend your effort.
My take
If you have been putting resources into getting mentioned in AI answers, you need to know that this channel is less stable than traditional search. The engines are still learning. Your strategy should be to keep an eye on industry trends, not to overhaul everything based on a single week of data. Measure against your competitors, not against yourself alone. That comparison is the only way to know whether a drop in your numbers reflects a real problem or just the engine changing how it works.
Sources
- AI search volatility: What it means for your traffic - HubSpot, November 20, 2026
They did the reporting. The writing and the opinions here are ours.
Questions people ask about this
What is AI search volatility
It is the unpredictable change in which businesses and websites answer engines like ChatGPT and Perplexity mention and cite from one day to the next. Unlike Google ranking changes, which you can track in a rank tracker, AI mentions shift in ways that are harder to measure consistently.
Why do AI answers keep changing
Language models are built to produce varied answers rather than identical ones. When researchers asked the same question twice in late 2025, ChatGPT and Google's AI system returned the same list of brands less than once in a hundred tries.
How do I know if my AI traffic dropped because of the engine or my content
Compare your traffic drop to what happened across your industry during the same period. If competitors also saw a drop, the engine changed. If only you dropped, the problem is likely your content or website structure.
Should I stop trying to get into AI answers
That depends on your resources and goals. This channel is less predictable than Google search right now. Keep measuring against your competitors and industry trends rather than expecting consistent results week to week.
What content types do AI engines prefer right now
Answer engines cite blogs, comparison pages, guides, and documentation at different rates. Watch which formats are growing in AI answers week after week and consider investing more in those formats.
This piece is about AI search volatility: What it means for your traffic, published by HubSpot. Go read the original; the opinions here are ours and the reporting is theirs.