An AI readiness assessment checks whether one workflow has the conditions a first AI project needs: a clear task, usable information, a named owner, explicit boundaries and a way to review the result. It is for a department or a whole company choosing where to start. You get a use-case brief, the gaps to close and a recommended next decision.
How it starts: one workflow, inside a department map of about 7 days or a company map of about 30 days. Fees, scope, timing, AI providers and data-handling terms are agreed on a short call before work begins.
What is an AI readiness assessment?
Here it means examining the prerequisites for one scoped AI project, grounded in the duties, tasks and software your team already uses. The same framework works for a department or a whole company, without pretending a whole company can be reduced to one score.
It does not establish legal compliance, approve system access or certify an organization as safe for autonomous operations. A free quiz can start a discussion, but it cannot verify the workflow, inspect the input data or establish who may authorize a change.
Does our business have to be AI-ready everywhere first?
No. Your business does not need to be ready in every department before making one useful improvement. It needs a well-defined use case and a realistic view of the conditions required to test it. Most companies start with a single task inside one department.
Which five questions does the assessment answer?
Is the task clear? Identify the trigger, input, action and expected result. "Help operations" is not specific enough to test.
Is the information usable? Establish where the input comes from, whether it is accessible for the proposed use and what happens when it is incomplete or inconsistent.
Is there a responsible owner? Name the person who can decide whether the project is useful and the person who checks the output. Performing a task and authorizing a change are different responsibilities.
Are the boundaries explicit? Decide what the AI may draft, suggest or change, and what must remain outside its scope. A task description does not grant credentials or permission.
Can the result be evaluated? Select a baseline and a review method. Record important errors, review effort and downstream effects, not only the time taken to generate an answer.
Why does the assessment begin with team discovery?
A manager may believe a task is predictable. The employee doing it may explain that a small group of customers always follows a different route. That exception can change the scope, the required data and the human checkpoint.
DutyGraph starts the readiness conversation with the work itself, using private, context-aware discovery to surface those details. The accounts still need review: an AI-generated task card is not a verified description merely because its fields are filled in.
Example: an internal request assistant
Illustrative scenario: An operations team wants AI to summarize incoming requests. A limited first project might draft a summary while a person chooses the response. Readiness questions include what request data may be processed, how unclear requests are flagged, how corrections are recorded and whether the assistant is kept out of approval decisions.
That team might be ready to test summarization without being ready to let an agent approve purchases or change a customer record.
What does an AI opportunity assessment add?
An AI opportunity assessment ranks candidate tasks before you choose one. It asks which tasks are repetitive, which have reliable inputs and which still need human judgment. Readiness then asks whether the task you chose can actually be tested safely.
If you are still choosing which tasks to examine, begin with the AI opportunity review inside the business operations audit. If the workflow itself is not understood, use process mapping services.
What does the assessment produce?
Agree a use-case brief, a list of prerequisites and gaps, explicit human review boundaries and a proposed next decision: test, narrow the scope, improve the inputs or defer. Where specialist security, legal or technical review is required, the brief names that dependency instead of treating a business assessment as a substitute.
How much does an AI readiness assessment cost?
The scope and the price are agreed on a short call before anything starts. The factors are the number of workflows, the number of participants and the evidence already available. See how pricing works.
Common questions
Is this the same as AI governance readiness?
No. Business-use-case readiness and governance readiness overlap but are not interchangeable. The AI governance readiness checklist covers the broader ownership and control questions; this page covers the operating conditions of one task.
Do you recommend specific AI tools?
We are not tied to one AI company. Tools such as Claude, ChatGPT, Cursor, GitHub Copilot, Gemini and any MCP client can read your map or take your agent designs. The assessment recommends what a task needs, then your team chooses the tool.
Will you tell us how much money AI will save?
No. The assessment states the hours a task takes today, from the team's own answers, and the conditions a test would need. A predicted saving is a prediction until your team measures it.
Can you assess a workflow we have not documented?
Yes. The assessment starts from what people say they do, not from documentation. Missing documents are recorded as gaps rather than filled in with assumptions.
Who stays responsible for the AI's output?
A named person, for every task. DutyGraph records the owner and the human checkpoint beside the task. The design is to keep delegated work inside approved human authority; enforcing permissions needs your own identity and access systems.
