Kedge
When the enterprise is stuck in dead calm, you don't drop an anchor to hold position — you cast a kedge far ahead and haul the whole ship forward.
Kedge finds the leader who can pull the organisation into its future — objectively, without favouritism, visibility bias, or pedigree halo.
The board already has a shortlist. That's the problem.
Before the engine sees a word, identity is stripped.
Names, pronouns, tenure, pedigree and org context never reach the model. Watch a raw intake memo pass through the Sanitization Service.
- Names & pronouns → tokens
- Pedigree → generalised
- Dates → tenure band
- Org context → functional mask
Zero-Knowledge by design: in this demo the candidate text never leaves your device.
Eight candidates. No names. No halos.
The engine scores strategic horizon (from authored artifacts), crisis performance (from simulation), and anonymised 360 feedback. Identity stays sealed.
Capability, mapped against what the future seat demands.
Objective readiness vs. the board's gut.
Same people. Stripped of visibility, pedigree and tenure halo — the ranking inverts.
Did we just trade one bias for another?
Kedge watches its own output. It compares the slate's composition before and after, and flags concentration the moment it appears.