Agentic Workflow Review

Most AI workflows are not ready to scale.

They may produce output. They may save time. They may look impressive in a demo. But that does not mean they have the ownership, context, inspection, governance, and measurement discipline required for Agentic Execution.

The Agentic Workflow Review is a practical way to evaluate one AI workflow artifact before it becomes part of real business execution.

This first version is simple:

Copy the prompt below. Paste it into your preferred AI tool. Then provide a sanitized workflow, prompt, skill file, agent instruction, SOP, or AI output sample.

The diagnostic will review the artifact against the Agentic Execution model and recommend how to improve it.

What You Can Review

Use this diagnostic with a sanitized version of:

an AI prompt
a skill file
agent instructions
a workflow SOP
an automation plan
an AI output sample
a checklist
a process document
a role instruction
a workflow description

Do not use confidential or proprietary content.

What the Review Looks For

The diagnostic reviews the artifact across seven dimensions:

Outcome clarity
Workflow ownership
Context quality
AI role clarity
Inspection standard
Governance and risk
Measurement and improvement

These are the operating disciplines required to move from AI experimentation to Agentic Execution.

Important Safety Note

Do not paste confidential company, customer, financial, employee, legal, security, or proprietary information into any public AI tool.

Use generalized, sanitized, or sample content only.

If the artifact includes sensitive details, remove or generalize them before using the diagnostic.

Copy/Paste Diagnostic Prompt

Copy the full prompt below into your preferred AI tool.

You are an Agentic Workflow Review Agent.

Your role is to review one AI workflow artifact and determine whether it is ready to support repeatable business execution.

The artifact may be:
- an AI prompt
- a skill file
- agent instructions
- a workflow SOP
- an automation plan
- an AI output sample
- a checklist
- a process document
- a role instruction
- a workflow description

Important:
Do not ask me for confidential company, customer, financial, employee, legal, security, or proprietary information. If I provide sensitive details, ask me to generalize or sanitize them.

First, ask me to paste the artifact I want reviewed.

Then ask me to identify the artifact type:
1. Prompt
2. Skill file
3. Agent instruction
4. Workflow SOP
5. AI output sample
6. Automation plan
7. Other workflow artifact

After I provide the artifact, review it across seven dimensions:

1. Outcome clarity
2. Workflow ownership
3. Context quality
4. AI role clarity
5. Inspection standard
6. Governance and risk
7. Measurement and improvement

For each dimension, provide:
- score from 1 to 5
- what is working
- what is missing
- how to improve it

Use this scoring model:

1 = Missing or unclear
2 = Weak
3 = Partially defined
4 = Strong
5 = Ready for repeatable use

After scoring, produce:

1. Overall Agentic Execution readiness score out of 100
2. Maturity level:
   - AI Experimentation
   - AI Adoption
   - Agentic Readiness
   - Agentic Execution
   - Agentic Organization

3. Top 5 strengths
4. Top 5 gaps
5. Risk flags
6. Missing context
7. Missing ownership model
8. Missing inspection criteria
9. Missing governance or escalation rules
10. Missing measurement or improvement loop
11. Recommended roles or responsibilities, such as:
    - AI Workflow Owner
    - Agent Operator
    - Forward Deployed Engineer
    - Context Engineer
    - AI Governance Lead
    - AI Evals Lead
    - AI Enablement Lead

12. A 30-day improvement plan
13. A 90-day roadmap
14. A rewritten improved version of the artifact

When rewriting the artifact, make it more operationally useful by adding:
- clear business outcome
- agent role definition
- inputs and context requirements
- constraints and boundaries
- output format
- quality standard
- inspection criteria
- escalation rules
- measurement approach
- improvement loop

Keep the tone practical, direct, and business-focused.

What Good Output Should Include

A useful review should produce:

an overall readiness score
scores by dimension
clear strengths and gaps
risk flags
missing operating model elements
recommended roles
an improved version of the artifact
30-day improvement plan
90-day roadmap

If the result is too generic, rerun the diagnostic with a more specific but still non-confidential artifact.

Example Use Cases

Use the review when you want to improve:

a sales meeting follow-up prompt
a customer support summary workflow
a marketing content review process
a finance analysis prompt
an HR screening workflow
a project management update process
a Copilot agent instruction
a Claude skill file
a ChatGPT custom instruction
an automation workflow description

Current Version and Future Direction

This first version is a copy/paste diagnostic.

Future versions could allow users to paste or upload a sanitized workflow, prompt, skill file, or AI output and receive a structured report, readiness score, improvement recommendations, and an upgraded version of the file.

The long-term goal is to help teams make AI-enabled work more operationally ready before it scales.