The most common question I get isn’t about any specific prediction — it’s about process. How do you actually do this? Where does the data come from? What makes a qualitative signal different from a guess?
Fair questions. Here’s how it works.
The Sources
We monitor 14 distinct indicator categories across four domains: investor communications (earnings calls, analyst notes, fund manager letters), practitioner channels (developer forums, internal corporate communications that become public, procurement discussion boards), media coverage (not just the headlines, but the framing, the sourcing patterns, and the editorial priorities), and policy signals (regulatory filings, legislative testimony, international governance frameworks).
None of these are secret. All of them are public or semi-public. The edge isn’t access — it’s attention. Most analysts are looking at the same data and ignoring the qualitative layer entirely because it doesn’t fit into a spreadsheet.
What We Track
For each indicator, we’re not measuring volume or polarity (positive vs. negative). We’re measuring conviction, specificity, and texture. A thousand people saying “AI is great” tells you nothing. The same thousand people, saying it with decreasing specificity and increasing defensiveness, tells you everything.
The detailed methodology is available in our free white paper. But the short version is: we read a lot, we read carefully, and we pay attention to how things are said, not just what’s said.
