About
A small company that grades itself in public.
One product, one bar: publish only what a circle of independent experts agrees on, then stand next to the result every Friday.
The company
The inspiration is unapologetically cinematic.
In Minority Report, three precogs float in a pool, and the system only acts when they agree — the interesting machinery is not the visions, it is the consensus rule. Strip away the fiction and that rule is just good epistemics: independent judges, forced agreement, and a record of when the room was wrong.
Apex Predictive is that rule built as software. A circle of independent expert models works every question against the live record; a convergence gate decides what survives; the board takes the survivors public; Friday grades them. No oracle, no vibes — a process that is allowed to say nothing, attached to a scoreboard it cannot edit.
The company is Apex Predictive LLC, a Texas limited liability company based in Keller, Texas — a sister company to Verumdare LLC, whose entity-intelligence platform feeds every deliberation.
Charles Sieg
Founder
Charles has been building software since 1997. In 2008 he founded Renkara Media Group, Inc. in Chicago and shipped AccelaStudy on the App Store's opening day — the store's first language flashcards app. Renkara grew to more than 300 published apps and 30 million downloads, then spent the following decade turning into an AI and machine-learning engineering company.
That second act produced AVIAN, an adaptive learning engine, and a patent portfolio he is the sole inventor on — 49 filings and 1,158 claims across 302 distinct inventions. He is also a cloud architect who has built platforms and infrastructure for Fortune 1000 companies.
Apex Predictive applies the same working method to a different bet: very few people, models running everything a model can run, and human judgment reserved for the questions, the gate and the grading.
AI-first, and not as a slogan
The experts are models; the research is model work over live data; the drafting is model work too. Human judgment goes where it compounds: what to ask, where the gate sits, and how a week gets graded.
The record is the pitch
No cherry-picked wins, no invented accuracy figures before a live record exists. The board's history, misses included, is the only performance claim we will ever make.
Resolvable or not published
Every statement names its source of truth and its deadline before it posts. If a question cannot be judged cleanly, it does not exist.
Skin in the game, eventually
The roadmap runs toward the engine backing its own convictions on regulated venues. We will get there on a public record, not on a promise.
Get in touch
Desks, funds, partners and press. This form is the way in.
Start a conversation
Tell us who you are and what you are trying to do. We answer everything.
We reply from a person, not a sequence. You will get one confirmation email to check the address is yours, and nothing else unless you ask for it.