iQueue.ai

How it works

One loop. Five layers.

A signal arrives. A team forms around it. Members reason with memory and skills, inside their authority. Decisions are recorded. The books get sharper. Then it happens again, without anyone asking.

The loop

Signal, team, work, decision, memory.

Analysts watching a live operational dashboard together
  1. 01

    Signal on the mesh

  2. 02

    Team self-composes

  3. 03

    Reason with memory

  4. 04

    Decide or escalate

  5. 05

    Record and learn

The layers

What the substrate actually is.

01

Agent runtime

Each team member is an agent with a role, a goal, a backstory and a library of skills it loads on demand.

Teams are crews. Crews can call other crews, so a plan built by one team can be interrogated by another while it is still being written.

02

Event mesh

A Redis-backed event bus. Everything that happens in the business is a signal: a data feed, a system event, a human decision, another team's output.

Teams subscribe to signals and wake when something relevant happens. That is what ambient means mechanically — no schedule, no click.

03

Semantic membranes

A governance boundary around every team and every party. A membrane decides what enters a team's attention, what leaves it, and in what form: it distills, redacts and routes.

It is judgment, not a filter — the difference between what a good employee thinks and what they say.

This is how a supplier's agent sits in a retailer's planning session without the supplier's cost data crossing the boundary, and how medical and commercial stay separated in pharma.

04

Memory fabric

Six layers: working, session, episodic, domain-operational, institutional, and federated — cross-organization and membrane-governed.

Before a team reasons or acts it retrieves a structured context pack: facts, constraints, feasible actions and open questions.

Mistakes are written back. The team never re-learns your business.

05

Decision authority and the ledger

Each member carries a decision matrix: decide alone, decide with a peer, or propose to a human. Escalation rules are explicit.

The decision ledger records every decision, its inputs, its reasoning and its outcome.

The ledger is the performance review, the audit trail, and the source of the books that set prices.

Model routing

The cheapest model that still hits the guaranteed outcome.

Every job is routed to the cheapest model that achieves the result we guaranteed. Outcomes are graded and fed back into routing. You never pick a model, and you never pay for a model you did not need.

Interoperability

The team is the integration point, not a database.

Teams work across your existing systems — CRM, ERP, ChMS, planning tools, Microsoft 365 — through governed connectors. We do not require a single system of record, because the problems we take on never had one.

How a team is onboarded

  1. 01

    Data-readiness assessment

  2. 02

    Role and authority definition

  3. 03

    Resumes approved

  4. 04

    Team goes live with recommendations only

  5. 05

    Autonomy expands as the ledger earns it

Human-in-the-loop by design

  • Members pause and ask when a decision sits above their authority.
  • Humans manage exceptions, not tasks.
  • Every action is reversible or approvable, per the authority matrix.

The guarantee

Everyone logs agent actions. We keep the books. Logs can't set a guaranteed price — books can.

Every competitor asks you to trust a demo. We ask you to read a contract.

Every digital worker is bonded

Each member of each team carries a bond. The team is accountable for the work, not just for the effort.

Fixed price, guaranteed result

Every job sits on a price list with a stated result. No seats. No token meters. If a job fails, we pay.

We keep the books

For every job we have ever run we hold what it cost, which model ran it, and whether it worked — pooled across thousands of organizations.

Hire the team. Keep the plan alive.