Process Readiness Assessment
Make the most of your assessment.
Before you start, this page covers how to define the unit of assessment, who should complete it and why process readiness matters. Reviewing that guidance will help you complete the assessment with a more accurate, defensible result.
Score the readiness of your selected process.
Answer against the process as it operates today, using evidence wherever possible. Score the current condition, not a planned future state and not the maturity of the organisation as a whole. Your result appears as soon as you finish. The value is in an accurate baseline, not a high score.
Loading the assessment.
How useful was this assessment?
Thank you. This goes straight into the next revision of the assessment.
Assess one process or value stream at a time.
Readiness is contextual. A well-documented, low-risk internal process may be ready for a controlled agentic pilot while a customer-facing or regulated process in the same organisation is not.
Complete this assessment for one clearly defined process or value stream. If you are considering several agentic use cases, assess each process separately. Do not average fundamentally different processes into one set of answers. Doing so conceals the very gaps the assessment is meant to expose.
Agree the context that makes your answers meaningful.
If this context has not been agreed, first run the Organisation Readiness Assessment with a multidisciplinary team. For the most useful result here, identify the following.
- The process or value stream being assessed.
- Its start, its end, and its material variations.
- The business outcome agentic AI is expected to improve.
- The decisions or actions an agent may be expected to take.
- The systems, tools, and data it would read or change.
- The process owner and the accountable executive.
- The people who should validate the answers.
- Relevant legal, regulatory, contractual, policy, security, and ethical constraints.
- The level of human involvement and oversight currently envisaged.
The screen does not infer readiness from enthusiasm, pilot activity or ownership of technology. It asks whether the agent is needed, what it may do, what could happen if it is wrong, whether people can intervene, and whether the organisation can contain and recover from failure. Seven questions are stop conditions: information fitness, proposed authority, consequence and reversibility, legal applicability, agent identity, human oversight, and containment. If one of those is absent or unknown, a high total cannot produce a positive result.
Adding an agent does not repair an unclear process.
AI agents can move work across functions and systems, but they inherit the conditions they meet. An undocumented exception becomes an unpredictable decision. An unclear owner becomes an accountability gap. Poor data becomes an incorrect action. An ineffective process becomes a faster ineffective process.
Deloitte's 2026 research found that only 5% of organisations regarded their business processes as highly prepared for AI agents. Many were layering agents onto existing processes even though only one in five considered their organisations prepared to redesign processes for autonomous operation. This assessment helps identify whether the selected process has the foundations agentic AI needs, and where discovery, redesign, or stronger controls should come first.
Deloitte, AI Agents Are Only the Beginning, August 2026 (501 US respondents in organisations already piloting agentic AI).
Five sections, 20 questions.
The screen covers 5 sections: Purpose and Process Suitability, Information and Technical Feasibility, Authority, Consequence and Regulation, Security and Control Feasibility, Value, Ownership and Change Capacity. Each question is answered against the process and use case as they stand today, with seven stop conditions that end the screen early. Work through them in the questionnaire itself.
Start with the process owner, then validate across the team.
One informed participant can produce an initial baseline. A decision to proceed should not rest on one perspective. Where practical, validate the answers with the people below. Differences between participants should be investigated, not averaged away: disagreement often signals unclear ownership, inconsistent performance, or undocumented practice.
- The accountable process owner.
- An operational subject-matter expert.
- Data and technology owners.
- Risk, compliance, security, privacy, or legal specialists relevant to the process.
- A representative of the people who perform or receive the work.
- The executive sponsor for the intended business outcome.
An actionable baseline for the process, not just a score.
The result is a structured readiness baseline. It is not certification, assurance, or authorisation to deploy an AI agent.
Your instant report includes
- Your next step: one of four outcomes.
- A screening score, used for triage.
- The stop conditions that are red or amber.
- A score for each of the five sections.
- A prioritised list of next actions.
- A copy of the result by email.
Use the result to decide the next step
- Clarify the process, outcome or proposed agent role.
- Redesign or simplify the process before adding an agent.
- Resolve a stop condition: authority, oversight, identity or safe fallback.
- Data-quality, lineage or access work.
- Legal, regulatory or security review for the use case.
- Secure the sponsor, participants or workforce capacity.
- Progress to the facilitated detailed assessment.
See the decision support you receive.
The sample shows the next step, the screening score, the stop conditions, the section scores and the next actions. The stop conditions decide whether the use case can progress; the score is only for triage.
This is an illustrative result. Complete the 20 questions for one defined process and use case to see your own.
One of four honest next steps.
The score is used for triage. Stop conditions decide whether a use case can progress. Consider both readiness and consequence: the greater the potential impact of an agent's actions, the stronger the evidence, controls, oversight and recovery arrangements should be.
If the process, intended outcome or agent role is still too vague to answer the questions against, that is the first thing to fix. Define them with the process owner and the accountable executive, then screen. The screen itself returns one of the four steps below.
-
Do not progress in its current form
A stop condition is absent or unknown. Reject, restrict or reframe the use case before further investment.
-
Foundation work required
The use case may have value, but important process, information, control, ownership or workforce conditions must be strengthened first.
-
Potential candidate, investigate
The use case merits further evidence gathering and specialist review before a detailed assessment.
-
Progress to detailed assessment
The use case is defined well enough to justify the facilitated detailed assessment. This does not mean it is ready to deploy.
The practical questions.
Everything you need to know before you spend 10 minutes on this.
- How long does the screening take?
- About 10 minutes. There are 20 multiple-choice questions across 5 sections, with no free text to write.
- What do I get at the end?
- An on-screen next step (do not progress, foundation work, investigate, or progress to a detailed assessment), a screening score used for triage, the stop conditions that are red or amber, section scores, and a prioritised set of next actions. A copy is emailed to you.
- Does the result mean we can deploy an agent?
- No. The result is a structured screening baseline, not a certification, technical assurance, or authorisation to deploy. It shows whether the use case is credible and bounded enough to justify a detailed assessment, and where foundation work is needed first.
- What is the detailed assessment?
- A facilitated cross-functional assessment IGX runs with you. It evaluates the process and the proposed agent together across eleven dimensions, using evidence, critical gates and an indicative autonomy ceiling. It is the next step for a use case that passes screening.
- Should one person complete this?
- You can, and the result is still useful as an indication. Treat it as provisional until the answers are validated with the process owner, the accountable executive, and the risk, data, and technology leads who hold the evidence.
- What if we have several processes in mind?
- Screen each one separately. Readiness in one process does not imply readiness in another. The questions are written for a single named process and use case.
- Is it free?
- Yes. There is no charge and no card required. Your result appears on screen the moment you finish.
- What happens to my answers?
- We use your answers to generate your result and email you a link to it. The result is saved to a private, unguessable link for 30 days so you can share it with colleagues. We do not share your details with third parties, and there is no sales follow-up until you reply.
Progress only when the evidence justifies it.
If the use case progresses, the detailed assessment examines eleven dimensions, requires evidence alongside the answers, and applies critical gates. It is a facilitated cross-functional assessment IGX runs with you, and it produces a weighted maturity score, an evidence confidence index, gate status, a readiness decision and an indicative autonomy ceiling. If the screen exposes wider issues in strategy, governance, infrastructure, workforce or accountability, return to the Organisation Readiness Assessment.
Process readiness depends on documented processes and trustworthy data. If your result points that way, start with discovering your real processes, building a single source of truth, or querying your operating model in plain language.
Controls and visibility enable scale.
IBM's 2026 research found that organisations embedding control into their AI systems experienced 25% fewer incidents than those relying on manual governance. McKinsey's 2026 research found that active mitigation lagged awareness across nearly every AI-risk category, and that explicit ownership was associated with greater responsible-AI maturity. Readiness work is not an obstacle to AI value. It establishes the visibility, accountability, and operating conditions needed to scale with confidence.
IBM Institute for Business Value and Oxford Economics, June 2026. McKinsey, State of AI Trust in 2026.
The screening provides an indicative baseline based on the answers supplied. It does not constitute legal, regulatory, security, or technical advice, certification, assurance, or authorisation to deploy an AI system. Validate results with appropriate evidence and specialist review for the process, use case, sector, and jurisdictions involved.