Sources is one of my favorite ZeroRank features, because it shows what's working right now in AI search for the prompts we're tracking. Every domain and every URL in Sources is something ChatGPT and the other AI models actually used to answer those prompts. If you care about GEO (generative engine optimization) or AEO (answer engine optimization), this is where you see which pages earn citations, and why. In this walkthrough I use Obsidian as the example brand. My own app, Typer, isn't online anywhere except its website yet, so it doesn't have much data. 🧩 HOW PROMPTS CONNECT TO SOURCES Your prompts live in the Prompts area, grouped into topics. Everything in Sources (every domain, every URL) is what the AI used to answer those prompts. Change your prompts and the sources change too. 🎛️ STEP 1: SET YOUR FILTERS The data changes depending on: • Topics: filter by topic, country and tags • Models: which AI models you're looking at • Time frame • Brand: your own, or a competitor's, so you can see the data as if you were them Filtering by topic matters if your product or service has more than one ideal client. Connect topics to each client segment, and when you work on articles or pages you want cited in AI answers, filter by those topics. 🌐 STEP 2: THE DOMAINS VIEW Domains is the default view. The percentages can go above 100% when a source is used more than once in a single chat reply. Domains groups all the URLs from the same site. For example, the different iA Writer URLs get grouped under one domain, same with Obsidian, apps.apple.com, Reddit and YouTube. Click a domain to see its URLs, sorted by how often they're used. Use the tooltips. They tell you when ZeroRank last fetched the content from each URL, whether your chosen brand is mentioned, and the URL type. 🔗 STEP 3: THE URLS VIEW For each URL you see: • URL type: listicle, comparison, product page, homepage • Whether your chosen brand was mentioned • Brand mentions and competitors on the page. Hover the "+3" to see the full list.