Your business is ready for AI when you have at least one repetitive, well-understood workflow, the data and systems it touches can be reached by software, someone owns the outcome, and you've decided how much risk you'll accept. You don't need a data warehouse, an AI team or a big budget. This AI readiness assessment scores you on 20 questions across six areas and tells you what to do at your score, because "not ready" usually means "ready for something smaller first".

Set aside fifteen minutes. Answer honestly, ideally with one or two team leads in the room, since owners tend to overrate how well processes are documented. Then read the scoring bands and the advice for each.

How does this AI readiness test work?

For each of the 20 statements, score:

  • 2 if it's clearly true today
  • 1 if it's partly true, or true for some teams but not others
  • 0 if it's not true, or you don't know

Maximum score is 40. Keep a note of which section each low score falls in. The section totals tell you more than the overall number, because a business can be strong on data and weak on ownership, and that changes what you should do first.

Section 1: Is your process clear enough to automate? (questions 1-4)

AI follows instructions. If nobody can describe how the work is supposed to be done, the AI will learn your inconsistency.

  1. We can name at least three tasks that happen dozens of times a week and follow a recognizable pattern. Email triage, order entry, quote preparation, appointment booking, invoice processing and status updates all count.
  2. For at least one of those tasks, a new hire could learn it from a written guide or a recorded walkthrough. If the only documentation is "ask Priya", score 0 or 1.
  3. We know roughly how long that task takes and how often it goes wrong. Even a rough baseline counts. Without one you can't tell whether AI helped.
  4. The exceptions are known. You can list the common ways the task deviates from normal: the customer who always sends the wrong format, the supplier who invoices in two currencies.

If this section scores low, your first project is a mapping exercise, not a build. Our guide to choosing which processes to automate first walks you through a two-hour workshop that fixes it.

Section 2: Is your data usable? (questions 5-8)

You don't need perfect data. You need data that's reachable and roughly consistent for the workflow you're starting with.

  1. The information the task needs lives in systems or shared files, not in people's heads or personal inboxes.
  2. Key records have consistent identifiers. Customers have IDs or unique emails, products have SKUs, orders have numbers. Duplicate customer records are the most common blocker here.
  3. We have examples of past work to learn from or test against. Past tickets and replies, processed invoices, completed quotes. These become your test set.
  4. We know where sensitive data lives (customer personal data, health or financial details, employee records) and who can see it.

Section 3: Can your systems connect? (questions 9-12)

An AI agent acts through tools. If your systems can't be reached by software, the AI can only advise, not do.

  1. Our core systems (CRM, accounting, ERP, helpdesk, booking tool) have APIs or ready-made integrations. Most cloud software does. Older desktop software often doesn't.
  2. We can get API access or admin credentials without a months-long vendor negotiation.
  3. Our main communication channels are in tools software can read, such as business email, WhatsApp Business, a helpdesk or a website chat, not personal phones.
  4. Someone on the team (or a partner) can set up and maintain integrations.

Low scores here don't rule AI out, but they change the shape of the first project. Our guide to connecting AI agents to your CRM and ERP covers the options, including what to do with systems that have no API.

Section 4: Are your people ready? (questions 13-16)

This is the section owners most often skip, and the one that most often sinks projects.

  1. There's a named person who will own the first AI workflow: review its output, report problems and decide when it's good enough.
  2. That person has time set aside for it. An owner with no hours available is not an owner.
  3. The team affected knows why AI is being introduced and what it means for their role. Rumors fill silence fast.
  4. Some people on the team already use AI tools for their own work and would help others get started.

Section 5: Have you decided how much risk you'll accept? (questions 17-18)

  1. We've agreed which decisions AI must never make alone (refunds above a limit, anything legal, medical or financial for an individual, anything sent to a regulator) and where a person must approve.
  2. We understand our privacy and industry obligations for the data involved. Depending on where you operate, that might include GDPR, India's DPDP Act, HIPAA or sector-specific rules. If you don't know, score 0 and check with counsel or your industry body. Our guide to AI data privacy and security covers the practical controls.

Section 6: Is there a budget and a reason? (questions 19-20)

  1. We can tie the first workflow to a business number: hours saved, faster response, fewer errors, more bookings, cash collected sooner.
  2. We've set aside a budget for a first project and its running costs, even a modest one, plus the review time it will need in the first months.

What does your AI readiness score mean?

Add up your scores.

ScoreWhat it meansWhat to do next
0-14Groundwork stage. AI would amplify confusion right now.Give staff an approved AI assistant with a usage policy, and spend a month documenting one process and its baseline.
15-24Ready for a contained first project.Pick one workflow with a clear owner. Start with an off-the-shelf tool or a simple automation with an AI step, plus human review on everything.
25-32Ready for a real AI workflow or a first agent.Scope a custom workflow or agent on your highest-value process. Plan for integration work and a review phase before loosening approvals.
33-40Ready to run several workflows.Build a short portfolio: two or three workflows, shared tools and monitoring, and an owner per workflow. Revisit governance quarterly.

Then look at your section totals. Any section scoring below half of its maximum is your bottleneck, regardless of the overall score. A business at 30 with a 1 out of 8 in "people" should fix ownership before building anything, because that's where the project will fail.

What do the common low-score patterns look like?

High systems, low process. Modern cloud stack, but every team does things differently. Classic in fast-growing companies. Fix: standardize one workflow before automating it.

High process, low systems. Clear, disciplined procedures running on a legacy desktop ERP or spreadsheets. Common in manufacturing and distribution. Fix: start with AI at the edges, such as reading inbound emails and documents into structured data, while planning system access.

High everything, low people. Strong data and systems, but nobody has time to own an AI workflow, or the team is anxious about it. Fix: free up an owner's time and have the conversation with the team first. Our guide on why teams don't use the AI tools you bought has scripts for it.

Low risk clarity. Everyone is keen, nobody has defined limits. Fix: one meeting to agree what AI may never do alone, written down before anything goes live.

How do you act on your score this month?

  1. Fix your weakest section first. Most fixes take days or weeks, not months: write a process down, get API credentials, name an owner.
  2. Pick one workflow that scored well on volume and pattern.
  3. Measure its baseline for two weeks.
  4. Plan a 90-day first project, from discovery to measured results. Our 90-day AI adoption roadmap lays it out week by week.

For the bigger picture of what AI can do once you're ready, see our no-hype guide to AI for business.

Frequently Asked Questions

What is an AI readiness assessment?

It's a structured check of whether your processes, data, systems, people, risk controls and budget can support an AI project. A good one tells you not just whether you're ready, but which specific gap to close first and what size of project fits you today.

Do we need a data team to be ready for AI?

No. Most SMB AI workflows run on the data you already have in your CRM, accounting tool, inbox and documents. You need the data to be reachable and reasonably consistent for one workflow, not a company-wide data platform.

We scored low. Should we wait before using AI?

Don't wait; start smaller. Give staff a safe, approved AI assistant, and use the time to document one process and measure it. Those steps cost little and move you into the next band within a month or two.

How often should we reassess AI readiness?

Retake it before each new AI project and at least twice a year. Scores change as you connect systems, document processes and build team confidence, and the bottleneck section often moves.

Can a consultant do this assessment for us?

Yes, and an outside view helps catch optimism about how documented your processes really are. But the questions here are designed so you can do a credible first pass yourself before paying anyone.

Want a second opinion on your score?

If you've taken the test and want help deciding what to do with the result, share your scores and your top workflow with us. We'll tell you honestly what size of project fits, including when the right move is groundwork rather than a build. See how we approach AI automation for small teams.