Start from the decision
Ask a question with an outcome, the one you actually bet on. Hunter builds everything else around it.
No data project first. You start where the money is.
AI model training, in your hands
Which artist will break through? Which company will make money? Train an AI to tell you.
Hunter trawls the public web for data points on thousands of examples, finds what actually predicts success, and trains a model for your question.
Give it something new and it gives you the odds, with the reasons behind them.
Your model
“Will this artist break through within 18 months?”
New artist
Nova Reyes
68%
Chance of breaking through · marker = 14% of all artists in the corpus
- ▲Listener growth four times the corpus median
- ▲Added to three editorial playlists this quarter
- ▼Only one release in the last twelve months
Sound familiar?
Your business is picking winners. Your method is instinct and a spreadsheet.
You see a hundred prospects for every one you can research properly.
Everyone watches the same headline numbers, so nobody gets an edge from them.
The best calls are gut calls, and gut calls are hard to defend to a partner.
You find out what predicted success after the deal has gone to someone else.
How it works
From a question to your own prediction model.
01
Ask a question
Pick an outcome you bet on. “Will this artist break through?” “Will this company be profitable?”
02
Name the data points
Tell Hunter what to collect: growth, releases, funding, headcount, press. Your expertise goes in first.
03
Trawl the public web
An agent searches public sources for thousands of examples and writes every value with its source.
04
Find what predicts
Hunter ranks the data points by how much they predict the outcome, and drops the noise.
05
Train your model
It trains a custom model on your corpus, then tests it on examples it hasn’t seen.
06
Get the odds
Give it something new. It gathers the same data points and returns the chance of the result you want.
The corpus
Thousands of examples. A source on every value.
Name the data points you care about. Hunter’s agent searches public sources, reads the pages it cites, and fills the corpus row by row.
When it can’t find a value, it says so. A model trained on guesses is worse than no model at all.
| Artist | Monthly listeners | 90-day growth | Editorial playlists | Releases (12 mo) | Tour dates |
|---|---|---|---|---|---|
| Halcyon Drift | 182kStreaming profile | +41%Streaming profile | 4Playlist pages | 3Discography | 22Venue listings |
| Mara Vey | 1.2mStreaming profile | +3%Streaming profile | 1Playlist pages | 1Discography | —Missing, not guessed |
| The Low Lanterns | 64kStreaming profile | +88%Streaming profile | 2Playlist pages | 4Discography | 31Venue listings |
2,400
Artists in the corpus.
14
Data points on each one.
91%
Of values found, each with its source. The rest marked missing.
Illustrative · fictional artists and values
What actually predicts success
It’s rarely what everyone is watching.
Everyone looks at the same headline numbers, so those numbers give nobody an edge. Hunter ranks every data point by how much it actually predicts the outcome.
In this example, momentum beats size. How fast an artist is growing matters six times more than how big they already are.
- 0190-day listener growth31%
- 02Editorial playlist adds22%
- 03Release cadence15%
- 04Tour dates booked12%
- 05Press mentions8%
- 06Total monthly listenerswhat everyone watches5%
- 07Social followerswhat everyone watches4%
- 08Signed to a labelwhat everyone watches3%
What it does
Your instinct, with the odds attached.
Your data points
Name the entity and its columns. Each is typed (number, date, link, text) and marked required or optional.
You decide what the model may look at. Your judgement shapes it from day one.
It trawls public sources
An AI agent searches the public web, reads the pages it cites, and fills your corpus row by row.
Weeks of research by hand become a hunt you start and come back to.
A source on every value
Each value keeps the page it came from. A value it can’t find is marked missing, never guessed.
The model learns from evidence, and you can check any number in it.
Find what actually predicts
Hunter ranks every data point by how much it predicts the outcome, and shows you the ranking.
You learn what matters. It’s rarely what everyone else is watching.
A model trained for you
Hunter trains a custom model on your corpus and your outcome. Not a general chatbot’s opinion.
An AI built for one question, and only yours.
The odds on anything new
Enter a new artist or company. Hunter gathers its data points and returns a percentage, with the factors that moved it.
A second opinion on every decision, in seconds, with its reasons.
Tested before you trust it
Hunter holds examples back, tests the model on them, and checks that a 70% comes true about 70% of the time.
You know how far to trust it before you bet on it.
Your edge stays yours
Your questions, corpus and models belong to your workspace.
Nobody else gets the model you trained.
Questions people train it on
If the outcome is public, you can train a model to predict it.
Investing
“Will this company be profitable within three years?”
Music
“Will this artist break through within 18 months?”
Sport
“Will this youth player make a first-team debut?”
Publishing and media
“Will this debut author earn out their advance?”
What it won’t do
Odds, not prophecy.
A model is only as honest as the data it learned from. So Hunter tells you where that data is weak.
- It gives odds, not certainties. A 68% means it will be wrong about a third of the time.
- Public data favours survivors. Success stories are everywhere and failures go quiet, so Hunter shows how its corpus was built.
- The past isn’t the future. A model learns yesterday’s patterns, so retrain it as the world moves.
- It isn’t for decisions about people’s jobs, credit, insurance, housing or education. We don’t sell it for those.
Request a demo
Forty-five minutes. Bring the bet you make most often.
We’re working with a small group of design partners on their own questions. The demo shows you how Hunter would build yours.
01
Your question
Which outcome you bet on, and what you look at today.
02
Your data points
Turn your instinct into the columns Hunter will collect.
03
The hunt
Watch Hunter build a corpus from public sources, with a source on every value.
04
What predicts
See how Hunter ranks the data points and drops the noise.
05
The odds
Score a new example and read the reasons behind its percentage.
06
Your questions
Ask anything, then decide whether to be one of our design partners.
Which bet would you make differently if it came with odds?
See Hunter turn your question into a sourced corpus, find what predicts success, and score something new.