Delta

Practical guides5 min read

An AI readiness checklist for your team: 12 questions to answer before you buy anything

A practical AI readiness checklist for your team: 12 questions on data, people, risk and value to answer before you spend on any tool.

Newan VinthusaFounder, Delta Camp
Illustration of a folder with a sheet of paper half pulled out, a pen resting across it and a small timer standing beside them.

The demo always looks good, which is exactly why it is a bad place to start. This AI readiness checklist is 12 questions your team can answer in one meeting, before anyone opens a procurement form. If you can't answer one, that is the finding, and it is far cheaper to learn it now than after the contract.

How to use this AI readiness checklist

Put the people who do the work in the room, not only the people who sign things off. Answer each question with a sentence, not a yes or no. Any question that takes more than a minute of arguing is where your real work is.

The 12 questions sit in four groups of three. The free AI Maturity Audit business track scores six areas (data foundations, processes, people, AI in practice, governance and risk, and strategy and value) in about 6 minutes. This checklist folds the same ground into four conversations you can have around a table.

Data and process: what the tool will touch

  1. Do you know where the information for this task lives, and who is allowed to open it?
  2. Could you write the task down as steps, with a start, an end and one named person responsible?
  3. Does the task come around often enough that a few saved minutes would show up in someone's week?

Question 3 matters more than it looks, because frequency beats size almost every time. I explain how to rank your tasks in audit before tools. Skip it and you end up automating the impressive task that happens twice a year.

Process is where teams trip. In a BCG survey of employees across Asia Pacific, only 57% said their companies are redesigning workflows to integrate AI effectively. A tool dropped onto an untouched workflow gives you the old way of working, only faster.

People: who will actually use it

  1. Do you know which AI tools your staff already use, including the ones nobody approved?
  2. Does a named person on the business side own this, rather than leaving it to IT alone?
  3. Can the people doing the work say what they would hand over and what they would keep?

The first of these tends to surprise people. The same BCG survey found that fifty-eight percent of APAC respondents say they would use AI even without company approval, and 35% would bypass restrictions to do so. Assume it is already happening and ask out loud, rather than guessing.

Ownership is the next one. Deloitte found that nearly 70% of Thai organisations see AI as primarily an IT responsibility rather than a cross-functional business priority. IT should be involved, but the person who feels the pain of the task should hold the pen.

Use cases come from the third. In the same Deloitte research on Thai organisations, 40% said they struggle to identify use cases. The people doing the work usually spot them faster than a steering group does.

Governance and risk: what could go wrong

  1. Have you flagged tasks that touch sensitive data, anything a customer sees before a person checks it, or a regulated process?
  2. Do you know who checks the output before it goes anywhere, and how long that check takes?
  3. Could you switch the automation off tomorrow, and does someone know how?

A flag does not mean don't automate. It means a person stays in the loop, and someone decides in writing how the output gets checked. The checking question is the one teams skip, and the four-hour tax is what happens when they do.

Strategy and value: what you want back

  1. Can you say, in hours per person per week, what you expect to get back?
  2. Have you decided where the freed time will go, so it doesn't drain into more email and meetings?
  3. Have you picked three tasks to start with, rather than fifteen?

More use does not automatically mean lighter work. PwC's Hopes and Fears survey of Singapore workers found that the share who used AI at work in the last 12 months rose from 69% in 2025 to 74% in 2026. In the same survey, 50% said AI has made their work more complex and 46% said it has added to their workload.

The second question in this group is the answer to that. I wrote about what happens without it in the capture gap. Time saved that nobody planned for simply disappears.

Reading your answers

Look at each group of three and mark where the answers were vague. A group with two vague answers is where your next few weeks should go, before any money goes on tools. Where the answers were crisp, you can start a small pilot.

Don't turn this into a score to defend in a steering meeting. The point is the conversation, and the silence in it. If you want to see where your answers place you, the four stages of AI maturity is the next read, or you can take the free AI Maturity Audit and get a stage and a first list of opportunities.

This week, put the first three questions in front of your team, answer them out loud, and write down which one got the longest silence.

Frequently asked questions

What is an AI readiness checklist?

It is a short list of questions about your data, processes, people, risks and goals that you answer before choosing an AI tool. The aim is to find the gaps while they are still cheap to fix. If a question gets a vague answer, that is where to start.

How do I know if my team is ready for AI?

Your team is ready when it can name a few repeating tasks, say who owns them, and say who checks the output. You don't need perfect data or a technical team. You do need people who can describe the work in steps.

Do we need technical staff before we start with AI?

Not to get a first win. The best early candidates are tasks that people already do every week, and the people doing them know the steps better than anyone. Technical help matters later, when you connect tools and set up governance.

Should we buy a tool first and work out the use case afterwards?

I wouldn't. A tool without a named task, an owner and a way to check the output tends to become a side window nobody opens. Pick three tasks first, then choose the tool that fits them.

How long does the AI Maturity Audit take?

The business track takes about 6 minutes. It scores six areas and lists your first opportunities, with a risk flag on each where it touches sensitive data, customer-facing work or a regulated process. It is free.

Where the numbers come from

Keep reading

Illustration of a desk with orange, green and grey sticky notes sorted into three columns, and a closed laptop pushed to one side.

Practical guides6 min read

Audit before tools: how to decide what to automate first

A 30-minute exercise for deciding which tasks to automate first: list your week, time it, and sort every task into three buckets before you buy an AI tool.

Read the post
Illustration of a printed draft covered in orange correction marks, next to an orange pen, a coffee cup and a desk clock.

Practical guides5 min read

The four-hour tax: the time you spend fixing what AI gives you

People save hours with AI, then spend most of them checking and correcting its output. How to spot the tasks that eat the gain, and what to automate instead.

Read the post