TL;DR: Before investing in AI, assess your AI Readiness. Use this 4-dimension framework and 12-question checklist to locate where your business stands and plan your next steps.

Is Your Business Ready for AI? An AI Readiness Checklist for SMEs

The question I hear most before advising any business is: "Are we ready?"

The answer is usually: you are readier than you think in some areas, and far less ready than you think in others.

This article provides a 4-dimensional framework for self-assessment—not just to get a "yes or no," but to identify what is missing and what to do next.

Quick Answer

The 4 dimensions of AI Readiness: (1) Data—do you have clean data for AI to work with? (2) Process—is your current workflow sufficiently standardized? (3) People—who will build and maintain the AI systems? (4) Budget—do you have enough to run experiments for 90 days? Many SMEs fail not from a lack of money or tools, but due to chaotic processes and unstructured data.

Why "Not Ready" Isn't a Money or Tool Problem

I have spoken with dozens of SME owners about AI over the past year. The most common pattern:

Company A: Has a budget, bought tools, hired consultants—yet 6 months later, nothing is running. Reason: Internal processes are chaotic, data is scattered everywhere, and no one knows exactly what problem they are trying to solve with AI.

Company B: Tight budget, no IT team, but has 3 stable workflows running after 8 weeks. Reason: They know exactly what problem they are solving, their processes are standardized, and they have one dedicated owner.

AI Readiness is not about scale or budget. It is about preparation in the right areas.

Dimension 1: Data Readiness — Do You Have the Data for AI to Work With?

AI is only as good as the data you feed it. Garbage in, garbage out—this rule is absolute when it comes to AI.

5 Questions to Assess Data Readiness:

1. Where is your operational data stored?

   • 🟢 Centralized in 1–2 tools (Google Workspace, Notion, CRM) → Highly ready
   • 🟡 Scattered but exportable → Needs 1–2 weeks of cleanup
   • 🔴 Mostly in people's heads or unorganized personal files → Data cleanup required first

2. Is your data consistently formatted?

   • 🟢 Standard templates exist; everyone follows the format
   • 🟡 Templates exist, but adherence is inconsistent
   • 🔴 No standard format; everyone does it their own way

3. Is your data regularly updated?

   • 🟢 Someone is responsible for updates; data is always current
   • 🟡 Irregular updates; occasional gaps
   • 🔴 Data is often stale; nobody knows what is current

4. Is there sensitive data that needs protection?

   • 🟢 Clearly classified: which data can go to AI and which cannot
   • 🟡 Unclassified; needs a review before granting AI access
   • 🔴 Highly sensitive data (customer, financial) lacks clear security policies

5. Do you have enough data volume for AI to recognize patterns?

   • 🟢 Enough examples (at least 20–30 cases) for each target task to automate
   • 🟡 Some examples exist, but more are needed
   • 🔴 Nothing for AI to reference

Interpreting Results: If you have 3+ 🔴 answers → Clean up your data before deploying AI. If you have 1–2 🔴 answers → Clean up data in parallel while starting with workflows that don't rely on that data. If you have mostly 🟢/🟡 → You are ready to proceed.

Dimension 2: Process Readiness — Are Your Processes Standardized?

AI cannot automate a chaotic process. It only accelerates the chaos.

3 Questions to Assess Process Readiness:

1. Is there at least one process you can describe step-by-step?

   • 🟢 Can write an SOP for it within 30 minutes
   • 🟡 Know the steps but have never documented them
   • 🔴 Done differently every time; no fixed process

2. Can the output of that process be clearly defined?

   • 🟢 Can describe "what a correct output looks like" in two sentences
   • 🟡 Know it when you see it, but hard to describe upfront
   • 🔴 Output varies wildly depending on context and recipient

3. Is the current process manual and repetitive?

   • 🟢 Done more than 5 times/week, highly repetitive, requires no complex judgment
   • 🟡 Done 2–5 times/week; occasionally requires judgment
   • 🔴 Done less than twice a week or always requires highly specialized expertise

The Rule: If you cannot describe a process clearly to a new hire, you cannot describe it to AI. Standardize first—automate second.

Dimension 3: People Readiness — Who Will Build and Maintain It?

This is the most overlooked dimension. Tools can be bought, and processes can be documented, but without dedicated owners, nothing sustains over the long run.

2 Questions to Assess People Readiness:

1. Is there someone ready to dedicate the first 20–30 hours to building the system?

   • 🟢 Yes, a specific person with dedicated bandwidth over the next 4–6 weeks
   • 🟡 Yes, but they are busy and can only spare 2–3 hours/week
   • 🔴 Everyone is at capacity; no one can prioritize this

2. Is there at least one person willing to learn and maintain it after setup?

   • 🟢 Yes, a tech-enthusiastic team member eager to own the system
   • 🟡 Yes, they will use it if ready but might not want to maintain it
   • 🔴 No one in the team wants to touch anything technical

Note: The builder doesn't need an IT background. They need the ability to read documentation, patience for trial-and-error, and a drive to solve problems. Most SMEs already have this person—usually the one who naturally optimizes their own daily work.

Dimension 4: Budget Readiness — Do You Have Enough to Experiment for 90 Days?

You don't need a massive budget. But you need enough to avoid stopping halfway.

2 Questions to Assess Budget Readiness:

1. Do you have a budget of 1–2 million VND/month for tools during the first 3 months?

   • 🟢 Yes, already approved
   • 🟡 Maybe, but requires justifying the ROI first
   • 🔴 No budget available for this at the moment

2. Can you commit one person's time for 90 days?

   • 🟢 Yes, can allocate 3–5 hours/week for them
   • 🟡 Possible but difficult; requires restructuring work
   • 🔴 Everyone is at 100% capacity

Interpreting the Results — Which Level Are You At?

If you are at Level 1–2: Don't buy more tools. Spend 2–4 weeks cleaning your data and standardizing at least one process first. This is a better investment than any AI subscription.

If you are at Level 3–4: Read the 90-Day AI Implementation Roadmap to get started right away.

Summary Checklist — 12 AI Readiness Questions

Data (5 questions):

• Centralized, non-scattered data
• Consistent data formatting
• Regularly updated
• Classified sensitive data
• Sufficient volume of examples for target tasks

Process (3 questions):

• At least one process described step-by-step
• Clearly defined output
• Process is repeated more than 5 times/week

People (2 questions):

• Has a builder to dedicate 20–30 hours to setup
• Has an owner to maintain it long-term

Budget (2 questions):

• Budget of 1–2 million VND/month for the first 3 months
• Can allocate 3–5 hours/week for the responsible person

If you want a tailored AI Readiness assessment for your business, message me on Zalo. I will walk you through these 12 questions and analyze the results to recommend your next best steps.

Read More

• AI for Vietnamese SMEs — Where to Start, What to Do, and How Much It Costs
• AI Implementation Costs for Vietnamese SMEs — Calculating ROI the Right Way
• What I'm Building and Why — A Founder's Real Perspective

FAQ

If my data is not clean, should I stop completely?

No need to stop completely. You can proceed in parallel: one person cleans up the data while another starts building a workflow that doesn't depend on historical data (e.g., writing marketing content, which doesn't require legacy data). The critical rule is not to feed dirty data into essential AI workflows.

Do early-stage startups (under 1 year) have AI Readiness?

Usually Level 2–3. The advantage: no legacy/outdated processes, making it easy to build things right from the start. The disadvantage: less historical data, unstable processes. Recommendation: standardize your processes first (even without AI), then introduce AI once things stabilize in 3–6 months.

Can family businesses or micro-merchants adopt this?

Yes, and it is often highly effective. Many micro-businesses have simple, highly repetitive processes where one person wears too many hats. AI can free up significant time. Start with the simplest use case: AI to answer frequently asked customer questions.

How long does it take to get "ready" from Level 1?

It depends on your current data and process maturity. Typically, it takes 3–6 weeks to clean up data and standardize 1–2 processes. After that, you can start your first workflow. The total timeline from Level 1 to having your first stable running workflow is usually 8–12 weeks.

💡 Related Articles:

• Claude Code Complete Guide: Installation and Real-world Application