The Agentic Enterprise  ·  Chapter ★
Chapter ★

Interactive Scorecard

40 questions · 4 pillars · 5 levels · share your results

40
questions
4
pillars
5
maturity levels
0
accounts needed

A self-assessment is not a certification. It is a mirror. The forty questions below are the ones we have asked, in some form, of every enterprise agent program we have walked through — and the answers, taken together, draw a picture clear enough to act on by the end of the morning.

Score each prompt on the five-point ladder: 1 Initial, 2 Repeatable, 3 Defined, 4 Managed, 5 Optimized. The anchor text on each level is deliberately short and behavioral — pick the rung whose description you can defend with evidence to a skeptical auditor, not the one you wish you were on.

Your answers are saved to this browser only. When you finish, you will see a four-pillar radar, a level badge, a gap heatmap, and a roadmap. A "Share this assessment" button encodes the whole thing into a URL hash you can paste anywhere — no account required.

How long this takes

About 20 minutes alone, or 90 minutes with three people in a room. We strongly suggest the second; the conversations are where the real assessment happens.

0 of 40 answered
Pillar · Governance

Governance

Pillar I — policy, accountability, and risk.

Accountable owner for the agent program
AI / agent acceptable-use policy
Model & agent inventory
Risk classification (impact × likelihood)
Alignment to NIST AI RMF or ISO/IEC 42001
EU AI Act / regional regulatory readiness
Incident response for agent failures
Bias, fairness, and impact assessment
Vendor & third-party AI risk
Workforce policy and disclosure
Pillar · Orchestration

Orchestration

Pillar II — frameworks, runtimes, evals, and guardrails.

Agent orchestration framework standardization
Memory architecture (short / long / episodic)
Evals & offline benchmarks
Tracing & observability
Guardrails (input + output + runtime policy)
Multi-agent coordination patterns
Cost & token-budget controls
Latency & reliability SLOs
Prompt & policy version control
Red-team / adversarial testing
Pillar · Use cases

Use cases

Pillar III — picking the right work, scoring it, scaling it.

Use case intake & qualification
Value modeling (ROI / payback)
Feasibility scoring (data, integration, model)
Human-in-the-loop design
User research & adoption metrics
Pilot → production graduation
Use case retirement
Domain expert involvement
Customer-facing vs internal balance
Lessons-learned & reuse
Pillar · Integration

Integration

Pillar IV — APIs, identity, data, and protocols.

Identity model for non-human actors
API & tool catalog
Data freshness & lineage
Vector / retrieval infrastructure
ERP / CRM / ITSM integration depth
Protocol adoption (MCP, A2A, OpenAPI)
Data classification & masking
Network egress & exfil controls
Logging, retention & audit
Disaster recovery & model failover

Your results

The chart and roadmap below update as you answer. They become useful around the 75% mark; before that, treat them as scaffolding.