AI readiness checklist for mid-size companies

Choose one workflow you'd like AI to improve, such as answering support questions, processing invoices or qualifying leads, and go through the six areas below. Count the items you can confidently check.

Most companies don't need to be ready everywhere. They need one workflow that's ready enough to prove value, and a plan for the gaps.

• The workflow happens often: daily or weekly, not once a quarter.
• You can name the metric it should move: time per case, backlog, response time, error rate.
• You know roughly what the work costs today in hours or money.
• One person owns the outcome and can make decisions.

• The information the work depends on exists in digital form.
• You know which source is correct when two disagree.
• You have examples of the work done well, to test against.
• You know which data is sensitive or regulated.

• The systems involved have APIs, or can get an API layer (see how to add AI to legacy apps).
• Someone can grant scoped, non-admin access for a project.
• You have a place to run it: an existing cloud account or a managed option.

• You know whether data may be sent to a public AI API, or must stay on private infrastructure.
• Spending limits and alerts exist in your cloud, per the guardrails checklist.
• You can log what an AI system reads and does.

• The team doing the work today is involved and sees the upside.
• It's clear which steps need human approval.
• Exceptions have somewhere to go when the AI isn't sure.

• There's budget for a build and for running it monthly (see what an agent costs).
• You've agreed what result in 90 days would justify continuing.

Our AI audit runs this assessment across your workflows in one to two weeks and ranks the opportunities by return and effort.