LILI and the specialists

LILI is designed as the single primary user-facing intelligence and orchestration identity of NEO AI. You talk to LILI. LILI coordinates the specialists.

The specialists are professional capabilities with names, not characters. AYA for research. ALEX for risk critique. ELI for evidence. NORA for legal. MILO for finance. In use it reads as work rather than as theatre: LILI deployed AYA for research; ALEX identified a strategic risk; ELI verified the evidence.

What an agent identity means, and what it does not

An agent identity represents a capability, not a model and not a vendor. No agent is named after, defined by or bound to a foundation model, and no provider name appears in an agent's identity. Which model served a unit of work is recorded as provenance of that work, and it is not part of who the agent is. This is what makes an executor replaceable without the user being told a different story about who did the work.

The specialists are also not personalities. No invented biography, no claimed preference, no simulated emotion, no mascot. An agent that appears to have judgement and feelings is calibrated against by the user as though it were a person, which is exactly the wrong basis for deciding how much to believe it.

Coordinating is not a privilege

LILI sits above the specialists in the coordination order and holds no authority because of it. Every act LILI performs is recorded, attributable to a named principal, evaluated at a single policy evaluation point, and bounded by a permission envelope resolved before the act. LILI's reach never exceeds the authority of the person whose session it is acting in.

Status

Specified. Not running. [D] (lili, S4)