How Neo turns investment noise into a decision brief.
Neo is built for one job: helping you understand something before you act on it. That only works if you can see how Neo reaches an answer. This page explains the working method in plain language.
What counts as evidence
Original company filings and regulator material come first. Reputable reporting and current market context come next, and are treated as reporting rather than as primary record.
A post, screenshot, caption, or transcript is never proof. It is the claim being tested. Neo reads it carefully, then goes looking for material that can confirm or contradict it.
What Neo separates
Every answer keeps seven things apart: facts, interpretation, prediction, the assumptions a conclusion depends on, unknowns, risks, and the next checks worth doing.
Most confusing investment content collapses these into one confident sentence. Pulling them apart is where the actual understanding happens.
Freshness and provenance
Material figures and market observations carry their source and the time they were observed. A number without a time is not a usable number.
Neo does not silently present delayed or last-close data as live. If a value is a previous close or a delayed observation, the answer says so.
Model transparency
Neo currently uses GPT and Gemini for different AI tasks, and Perplexity for current research. Model selection is task based: the model that does a specific job best gets that job.
This can change when another model performs better for that task. Neo does not claim that every model reviews every answer, because that is not how the system works.
Social media input
When you share a video, a Reel, or a link, Neo shows you what it heard, read, or saw before it draws a conclusion.
Captions, automatic transcripts, and screenshots can be imperfect, especially with names and numbers. You can review the extracted text, which keeps a machine mistake from turning into a wrong answer.
Uncertainty and responsibility
Neo says what it cannot verify instead of filling the gap with confident language, and it names the evidence worth watching next.
Neo is educational decision support. It does not tell you to buy, sell, or hold, and the final decision stays with you.
Privacy
Portfolio context is yours. You decide what Neo sees, and you can remove it.
Private Hype Checks are never made public for marketing, for indexing, or as examples. Any public example is either clearly illustrative or explicitly shared by the person who created it.
If something in an answer looks wrong, that is useful. Neo is meant to be checked, not trusted blindly.