Breadth across the organization. Depth into the architecture.
Diagnosing where you stand today to move up the levels by working through the build sequence and instrumenting the stack.
AI Readiness Levels.
A diagnostic that tells you where you stand today and what moving to the next level actually requires.
- L0Dormant
No meaningful AI use.
- L1Role-level lift
Individual usage, ungoverned. Gains stay with the person.
- L2Governed integration
One bounded workflow runs end to end under a control plane and leaves evidence.
- L3Coordinated hybrid
Several workflows run under shared policy, with task-node contracts carrying real work.
- L4Adaptive autonomy
Learning compounds across workflows and updates policy and memory.
Client takeaway · "We know where we are, what Level 2 requires, and what evidence proves we reached it."
The Intelligence Stack.
Seven layers, each requiring the one beneath it. The bottom four build the world model, layers five and six govern execution, and layer seven compounds.
Institutional memory, policy updates, and workflow learning. The system improves overnight.
Traces, evals, proofs, acceptance decisions, auditable rationale.
Agents, humans, tools, queues, workflows, APIs, systems of record.
Policies, authority routing, human-agent handoff rules, escalation paths.
Synthesis, anomaly detection, causal hypotheses, prioritization, confidence scoring.
Canonical identity, relationship graphs, temporal state, source provenance, context distillation.
The business made readable, with entities, workflows, and vocabulary named so a system can read them.
Client takeaway · "We know what needs to be built, and where the world model lives."