In 2010, "buy running shoes online" was a useful keyword to track. 20,000 searches a month. Clear intent. You build a landing page, you rank for it, you get traffic.
In 2024, the person who wants to buy running shoes is typing something like "I have wide feet and my last pair gave me blisters after 5k, what should I get for half marathons?" — into an AI chatbot. And the answer they get doesn't come with a list of blue links to click.
This shift is breaking the keyword-volume model of SEO and content marketing. And prompt tracking is what's replacing it.
What prompt tracking actually is
A "prompt" is a full, natural language question that a user inputs into an AI system. Unlike a two-word keyword, a prompt contains context, constraints, and intent.
Prompt tracking means monitoring what kinds of questions and requests people are using in your category. Which problems are they trying to solve? What comparisons are they making? What features are they asking about?
You can't get this data the way you get keyword volume data — nobody publishes prompt frequency numbers. But you can piece it together from a few sources:
- Reddit threads in your niche (look at the questions being asked in the comments)
- Your own customer support inbox
- "People Also Ask" boxes in Google (these are increasingly pulled from AI-style natural language queries)
- Q&A forums and Quora in your category
Why this changes how you make content
A keyword strategy optimizes for ranking. A prompt strategy optimizes for being the most useful, specific answer to a complex question.
If someone asks ChatGPT "what link shortener should I use if I need custom domains, UTM tracking, and team workspaces?" — and ChatGPT recommends you — that's a citation worth more than most page-one rankings. The person asking is highly specific about what they want. They're close to making a decision. And they trusted the AI enough to let it narrow the field.
To show up in answers like that, you need content that actually addresses those specific multi-part questions. Not an SEO article padded with generic information, but a specific, well-structured answer to a specific, complex question.
The link between prompt tracking and click attribution
When a user does eventually click through from an AI recommendation or a piece of content you created, you want to know about it. The referrer won't say "chatgpt.com/thread/xyz" in most cases — you'll see an in-app browser referrer or Direct traffic.
Create content with tracked links built in. If an AI cites your page and a user follows the link, that click goes through your tracker. Over time, you'll start to see which content pieces are generating the most AI-assisted clicks — and that tells you which prompt strategies are working.