The Growth Leader’s AI Readiness Audit: Are You Using AI to Scale Your System or Just Your Output?
A founder told me last month: “I run a lean business. I do not need an AI readiness audit. I need people to move faster, ship more campaigns, and cut agency cost now.”
Fair. Speed is a real gain. But it proves nothing about the growth system. Shipping more campaigns does not mean AI improved your decisions, your personalisation, your measurement, or your learning speed.
There is a difference between using AI to produce more output and using AI to build a better growth system. Most companies do the first and call it the second.
This article gives you the audit that separates the two. It runs across four dimensions:
- Data infrastructure
- Decision automation
- Personalisation capability
- Measurement integrity
Score yourself honestly on all four. You will know within an hour whether your AI adoption is strategic or cosmetic. My argument is simple: if AI only makes your team faster, you have bought productivity. You have not bought growth.
1. Adoption Is Rising Faster Than Maturity
Singapore SME AI adoption more than tripled, rising from 4.2% in 2023 to 14.5% in 2024, driven mainly by off-the-shelf generative AI tools. And 50% of Singapore businesses plan to prioritise AI over the next twelve months.
Now the uncomfortable numbers. 71.5% of Singapore firms had not begun AI adoption, and only 3.8% had integrated AI into core processes. Among firms already using AI, 70.7% report productivity gains, but only 13.3% report better decision-making.
That gap is the whole story. Teams feel faster. Decisions stay the same.
Using AI only to make more content is adding reps with poor form. You feel productive. You are scaling bad mechanics. The audit below checks your form first.
💡 Key Takeaway: Productivity gains are the floor of AI value, not the proof of it. The proof is better decisions.
2. Data Infrastructure: Can Your AI Reach the Truth?
Most founders put AI inside marketing before they put it inside data. That order is backwards.
An AI tool writing campaigns on top of messy customer data produces polished guesses. No model decides well when the inputs are wrong.
The global pattern backs this up. McKinsey found 88% of organisations use AI in at least one function, but only about one-third have begun scaling their AI programmes. The high performers redesign workflows, build data infrastructure, and embed AI into core processes.
Three audit questions:
- Is your first-party customer data unified and accessible?
- Can your AI tools query it directly, or do humans copy and paste?
- Do product, marketing, and finance read from the same numbers?
Two or more noes mean your AI runs on guesswork.
3. Decision Automation: Volume Is Not Judgement
Klarna automated hard. Its AI assistant handled 2.3 million conversations, two-thirds of its customer service chats, doing the work of 700 full-time agents. Then Klarna moved to restore human support options.
The lesson is not that automation failed. The lesson is that Klarna optimised for automation volume, and volume is not judgement.
So audit the decisions, not the tasks. Which calls did AI actually improve this quarter? Budget shifts, pricing moves, channel choices, retention plays. If the answer is none, you automated activity, not judgement.
💡 Key Takeaway: Count the decisions AI changed this quarter. Zero means you bought speed, not a system.
4. Personalisation: Better Decisions Per Customer, Not Better Copy
Most teams think personalisation means sharper email copy. It means making a better decision for each customer, at scale.
DBS shows the ceiling. The bank runs over 1,500 AI models across more than 370 use cases, generated over SGD 750 million in economic value in 2024, and delivered 1.2 billion personalised nudges to more than 13 million customers.
You do not need a bank’s budget. You need the same architecture: data feeding models, models feeding customer-level decisions, decisions feeding measured value.
The audit question: does your AI change what each customer sees, pays, or receives? Or does it reword the same message for everyone?
5. Measurement Integrity: If You Cannot Verify It, You Own the Risk
Air Canada learned this in court. A tribunal held the airline liable after its chatbot gave a passenger incorrect bereavement fare information. The AI spoke. The company paid.
Measurement integrity has two parts. Validation: someone checks what your AI tells customers before it costs you money. Attribution: you can trace AI activity to commercial results, not to vibes.
An AI stack without validation scales errors at machine speed. An AI stack without attribution cannot prove it grew anything.
💡 Key Takeaway: If you cannot measure what your AI decided and what it earned, you are not running a system. You are running an experiment with no control group.
Final Thoughts: If AI Only Makes You Faster, It Is Not Making You Better
The founder at the start wanted speed, and speed is fine. But speed compounds whatever system it runs on. A weak growth system with AI on top is a weak system producing more.
Run the audit. Score data infrastructure, decision automation, personalisation, and measurement integrity from one to five. Anything under three is where your AI spend is leaking.
The pattern I see most often: strong tools, weak system. Content output up, decision quality flat. That is cosmetic adoption, and the numbers expose it within two quarters.
If you want a second pair of eyes on your scores, book a discovery call or connect with me on LinkedIn. Bring your four scores. I will tell you which dimension to fix first, and why it is rarely the one you think.
A note before you close this tab. The fact that you read this far tells me something. You already sense that the way you’ve been thinking about growth might be incomplete. That instinct is worth following.
Mervyn Chua is a growth-transformation consultant helping founders and CEOs build the strategic clarity and systems to grow in an AI-first world. If this raises questions worth exploring for your brand, let’s talk.
