Agent Readiness Assessment
How ready is your organisation for AI agents?
Score your readiness across six capability dimensions.Get your stage, your score, and the highest-leverage move to make next.Built for any team bringing agents into production.
Live signal
The agent shift, measured every week.
Three live readings on the agent economy, who's hiring, which tools have real demand, and what just moved. They all point the same way: agents are reaching production faster than the controls around them.
2,027 agentic roles open across 213 companies 65% / 61% agentic share, frontier vs enterprise hiring 92 agent tools ranked by real-world demand
The Agent Brief
Read the brief → Google, Microsoft, Cisco, Nvidia, Salesforce launch ARD protocol for agent tool discovery and safety Researchers demonstrate SearchLeak prompt injection attack against M365 Copilot Enterprise US government forces Anthropic to pull Fable 5 and Mythos 5 models over guardrail bypass Developers seek reproducible execution patterns as Claude Code exhibits non-deterministic behavior
Heating up Agent safety and governance frameworks 4×Cursor IDE adoption and capabilities 4×Agent oversight and compliance standards newClaude Code reliability and safety issues 2×Enterprise AI compliance and standards 11×
Why it matters
One agent becomes infrastructure before anyone governs it.
A useful agent becomes a shared workflow. Then it gets system access. Then it touches customer data. By the time risk teams notice, no one can clearly answer who owns it, what it did, or how to stop it.
- It starts as one agent. Shipped by whoever was keenest, no inventory entry, a personal token, no review.
- Teams adopt it. It spreads across marketing, sales, support and finance, with fragmented ownership.
- It reaches production data. Now it's wired into the CRM, Drive and the warehouse, with no runtime visibility.
What good looks like
The five properties of a production-ready agent.
A trustworthy agent is inventoried, owned, observable at runtime, access-controlled, and continuously evaluated for drift. Without these five, you cannot measure trust, or manage risk.
Inventoried
Every agent is registered with a purpose, an owner, and a defined scope of access. Without a registry, you cannot answer "which agents are running?"
Agent registry →Owned
Each agent has a named human owner accountable for its behaviour, change management, incident response, and risk acceptance.
AI agent governance →Observable
Every action an agent takes, tools called, data accessed, decisions made, is logged, attributable, and visible in real time.
Agent observability →Access-controlled
Permissions follow least-privilege. Scopes can be granted, narrowed, and revoked without code changes, and agent identity is distinct from user identity.
Agent identity →Continuously evaluated
Quality, safety, and policy compliance are tested in production, not just before deployment. Drift and regressions are caught automatically.
Runtime governance →The model
Five stages, from experiment to production-ready.
The Agent Operational Maturity Model maps where you stand. Most enterprises sit between Stage 02 and Stage 03. Production-readiness lives at Stage 05.
- 01 Experimental Isolated agent use by individuals.Next control to fixStart an inventory, even a spreadsheet. You cannot govern what you cannot list.
- 02 Shared Agents reused across teams.Next control to fixAssign named owners. Move agents off personal tokens onto scoped identities.
- 03 Operational Agents touch systems, APIs, and data.Next control to fixAdd runtime visibility. Log every tool call, data access, and decision.
- 04 Governed Reviews, owners, audits, and evaluations exist.Next control to fixAdd runtime enforcement, policies that block, scope, or escalate in the moment.
- 05 Production-Ready Agents are monitored and controlled at runtime.Next control to fixMaintain. Quarterly evaluations, drift checks, and incident postmortems.
The assessment maps six capability dimensions
Strategy & ValueIs agent adoption tied to business goals, and do you measure the value?
Capability & AutonomyHow capable and autonomous are your agents, really?
Data & Tool ReadinessCan your agents reach the data and tools they need to act?
Orchestration & IntegrationAre agents embedded in real workflows, and orchestrated across them?
Evaluation & ObservabilityDo you measure whether your agents are doing a good job?
Operating Model & SkillsDo you have the team and the path to run agents at scale?
Readiness assessment
How ready is your organisation to build and run agents?
Twelve scenario questions across six capability dimensions, synthesised from the leading agentic maturity frameworks. You'll get a read on each dimension as you go, then your overall readiness stage, a score, and the highest-leverage move to make next. About four minutes.
Your readiness stage
Readiness score0/100
Biggest opportunity
Do this next
Answer engine
AI agent governance questions, answered directly.
Short definitions for teams evaluating AI agent governance, runtime control, and production readiness.
What is an AI agent readiness assessment?
An AI agent readiness assessment measures how prepared your organisation is to build, deploy and run AI agents in production. It scores six capability dimensions, strategy and value, capability and autonomy, data and tool readiness, orchestration and integration, evaluation and observability, and operating model and skills, then places you on a readiness ladder from Exploring to Agent-Ready with the highest-leverage move to make next.
How do I assess AI agent readiness?
Assess AI agent readiness by scoring six capability dimensions against a readiness ladder. Answer two scenario questions per dimension, total the points, and map your percentage score to a stage: Exploring (0 to 25), Experimenting (26 to 50), Operationalising (51 to 75), or Agent-Ready (76 to 100). The Agent Readiness Assessment on this page does this in about four minutes and returns your stage, score, weakest dimension and next step.
What are the dimensions of AI agent readiness?
The six dimensions of AI agent readiness are: Strategy and Value (use cases tied to business goals and measured value), Capability and Autonomy (how capable and autonomous your agents are), Data and Tool Readiness (clean data and scoped tools agents can act through), Orchestration and Integration (agents embedded in real workflows and orchestrated across them), Evaluation and Observability (measuring agent quality in production), and Operating Model and Skills (the team and path to run agents at scale).
What does it mean to be ready for AI agents?
Being ready for AI agents means you can move agents from pilot to production reliably: you have prioritised use cases tied to value, capable and appropriately autonomous agents, clean data and scoped tools, orchestration across workflows, continuous evaluation, and an operating model with named owners and a repeatable path to production. Most organisations sit in the early Exploring or Experimenting stages.
What is AI agent governance?
AI agent governance is the operating model for knowing which AI agents exist, who owns them, what they can access, how they are evaluated, what policies constrain them, and how their actions are audited and controlled in production.
What does agent operational maturity mean?
Agent operational maturity describes how ready an organization is to run AI agents as operational infrastructure. It includes inventory, ownership, runtime visibility, access control, evaluations, auditability, policy enforcement, incident response, and compliance readiness.
Why do AI agents become risky after deployment?
AI agents become risky after deployment because they can spread across teams, connect to systems, access sensitive data, affect workflows, and trigger decisions without the same controls used for other production infrastructure.
What does production-ready mean for AI agents?
A production-ready AI agent is observable, owned, evaluated, access-controlled, auditable, policy-constrained, and governed at runtime.
Who should take the Agent Readiness Assessment?
The Agent Readiness Assessment is designed for CTOs, CISOs, Heads of AI, enterprise architects, platform engineering leaders, security teams, governance leaders, risk teams, and AI transformation leaders.
How do I assess AI agent operational maturity?
You assess AI agent maturity by measuring six capability dimensions, strategy and value, capability and autonomy, data and tool readiness, orchestration and integration, evaluation and observability, and operating model and skills, against a maturity ladder from exploring to agent-ready. Most organisations are still in the early, experimental stages.
What controls do AI agents need before production?
Before production, AI agents need a registered inventory entry, a named owner, runtime observability, least-privilege access scopes, continuous evaluations, and an incident response path. Without these six controls, an agent in production is shadow infrastructure.
What is the difference between AI governance and AI agent governance?
AI governance covers models, datasets, and outputs. AI agent governance adds the operational layer: which agents exist, who owns them, what they can do at runtime, and how their actions are audited and controlled in production.
What is runtime governance for AI agents?
Runtime governance is the enforcement of policies at the moment an AI agent takes an action, blocking, scoping, escalating, or logging tool calls and data access in real time, rather than only reviewing behavior after the fact.
What is AI agent drift?
AI agent drift is the gradual change in an agent's behavior, accuracy, or safety profile over time, driven by model updates, changing tool responses, prompt context shifts, or upstream data changes. Without continuous evaluation, drift is invisible until something breaks.
Understand where your organisation really stands.
Stage 02 → 03 is where most enterprises sit today: agents spreading across teams and systems, with very few of the controls needed to scale them safely.
