Early enough to act. Light enough to adopt.
HumanGraph is a prediction layer on the stack you already run — reading each player from their first session, not their third month.
One flat export in, decision-grade fields back, landing in the CRM, BI and tools your teams already use. A layer on top — not a system to rip out and replace.
Your stack tells you what happened. The best of it tells you what's next — a month late.
Most stacks lean on what's already happened — reporting that looks backward, or prediction that waits for deposit history to pile up and refreshes overnight. Either way, a future high-value player looks average until the deposits land — and by the time anything catches up, the welcome window's already closed. HumanGraph reads the earliest behavioural footprint instead: the first session, before the history exists. Same player record, weeks earlier — and because it sits as a layer on top of what you already run, getting it earlier doesn't mean rebuilding anything.
It plugs into your stack. It doesn't rebuild it.
You send
Flat player-day tables. Nothing exotic, scoped to what you already maintain.
We process
The signals are generated under terms agreed before anything connects.
You get back
A small set of decision-grade fields on the player record — CRM · BI · automation.
The fields land where your teams already work — same CRM, same BI, same screens. No new dashboard, nothing to rip out.
Live in weeks, not quarters.
No data-warehouse rebuild. No months-long integration. The work is a flat export and a calibration window — here's the path.
Connect
You send a sample export; we confirm the fields map cleanly.
Calibrate
The signals are tuned to your player base and market. Launching from zero? Cold-Start priors stand in until your own data builds up.
Live
Fields start flowing onto the player record, into the tools your teams already use.
Validate
You measure the signals against your own outcomes before anything scales.
Our cloud, or your own environment.
Run the signals in our managed cloud, or deploy HumanGraph inside your own infrastructure — same signals either way. Pseudonymised inputs, no direct identifiers, every decision auditable. Operators in regulated markets usually run on-prem; you pick the deployment that clears your compliance bar.
See the full security & deployment detail →Already live, or starting from zero.
Pilot
Test the signals on your own data — validate quality, workflow fit and commercial impact before any broader rollout.
Explore a pilot →Cold-Start
Calibrate from regulated-market priors and your first weeks of real data — before your first promotional decisions.
Explore Cold-Start →Twenty minutes. One question: would earlier change your numbers?
Not a procurement cycle, not a platform commitment — just a conversation about whether operating earlier moves your metrics.