2/4
We tested whether Meridian could find costly developments before they became obvious.5 modeled companies.
2,000 signals.
5 planted developments.Meridian surfaced all 5 early.Revenue opportunity — 118 signals early
Client churn — 116
Margin leakage — 129
Delivery risk — 115
Blocked decision — 142 £825,000 in simulated exposure. Controlled testing. Not customer ROI.
Today we’re introducing Meridian - AI built to understand an entire company.
Meridian finds the problems, risks and opportunities your company is missing — before they become obvious.
In controlled testing, Meridian surfaced 5/5 costly developments early.
See what your company is missing:
meridiankernel.com/
Most company software tells you what happened.
We built Meridian to determine what is happening, why it is happening, and what requires a decision.
15 source classes. Evidence-linked reasoning. Human-governed action.
The AI situation engine for companies.
🤖 Made with AI
Today we're introducing Meridian, the AI Situation Engine for companies.
Meridian connects evidence across an entire company to identify what changed, why it matters, and what should happen next.
This is the beginning of companies that can understand themselves.
Most operational failures begin as ordinary records in different systems.
In a frozen synthetic benchmark, Meridian connected one developing risk at signal 173. Formal escalation arrived at 289.
A 116-signal window.
Can you beat it?
meridiankernel.com/break-mer…
173 signals. Eight systems. One fictional company.
Meridian reached one conclusion. The answer is already frozen.
Now everyone gets the same evidence and one job:
Find a better answer.
meridiankernel.com/break-mer…
Synthetic case. Failures will be published.
Our AI scored 100%.
I assumed the test was bullshit.
So we tried to break it.
We did.
Rebuilt it. Broke it again.
Now Meridian catches hidden developments 115–142 signals early in our synthetic tests.
Customers: 0.
Prove Meridian is bullshit. ↓






