AI Readiness
AI Readiness Checklist: Before You Automate Anything
Before choosing tools or building agents, check whether your people, processes, information and controls are actually ready.
7 min read
In this guide
Buying an AI tool is easy.
Making it useful inside a real organization is harder.
Before automating a process, check the basics.
1. Is the business problem clear?
Write the problem in one sentence.
Not:
“We need AI.”
Better:
“Our team spends too much time answering the same internal policy questions.”
If the problem cannot be explained clearly, the project is probably not ready.
2. Is the current process understood?
Before automating a process, understand how it works today.
Know:
- where it starts
- who is involved
- which systems are used
- where decisions happen
- where delays occur
- what exceptions exist
Automating a confusing process often creates faster confusion.
3. Is the information usable?
Check whether the information needed by the AI is:
- available
- accurate
- current
- organized
- owned by someone
- safe to use
Poor information creates poor outcomes, no matter how capable the model is.
4. Are the right systems accessible?
Ask what the AI or automation actually needs to connect to.
That might include:
- CRM
- documents
- internal knowledge
- forms
- spreadsheets
- business software
Do not assume an integration is possible until the relevant access, APIs and permissions have been checked.
5. Is ownership clear?
Someone should own the process.
Someone should also be responsible for:
- reviewing results
- reporting problems
- approving changes
- maintaining information
- deciding when human intervention is needed
“AI owns it” is not a governance model.
6. Are the risks understood?
Think about:
- confidential information
- customer data
- inaccurate outputs
- unauthorized actions
- access permissions
- decisions that need human approval
Higher-impact tasks usually need stronger controls.
7. Will people actually use it?
A technically good system still fails if it makes the team’s work harder.
Ask the people who do the work:
- What slows you down?
- What would genuinely help?
- What should never be automated?
- What would make you trust the system?
Adoption starts before deployment.
8. Can success be measured?
Choose a few useful signals.
Examples:
- time saved
- response time
- number of manual steps
- completion rate
- error/rework rate
- employee adoption
Measure what relates to the original problem.
Readiness check
You are in a stronger position to start when:
- the problem is clear
- the current process is understood
- useful information exists
- required access is possible
- ownership is assigned
- risks are known
- users are involved
- success can be measured
AI readiness is not about having the newest technology.
It is about being ready to use technology well.
Ai MindUp
Want a clearer starting point?
Our AI Readiness Assessment looks at the practical factors that shape successful AI adoption.