Nov 15, 2025
By
Arsedian Ivan
Executives are launching "AI transformation initiatives" as if AI exists separately from their digital evolution.
This is why so many AI projects fail, not because the technology doesn't work, but because you're trying to build skyscrapers on foundations made of sand.
There's no such thing as AI Transformation. There are just levels to Digital Transformation and AI is simply one stage in that journey.
My tips to C-Suite:
#1: Don't push AI transformation separate from your digital journey.
Creating parallel tracks, one team on "digital transformation," another on "AI transformation", creates silos and wasted effort. AI isn't a separate journey.
#2: You can't skip levels + Garbage In Garbage Out.
You can't jump from basic data collection (Level 1) straight to autonomous AI (Level 5). Yet organisations constantly try this, buying enterprise AI licences while their foundational data is a mess. Result? Garbage in, garbage out at machine speed.
#3: Learn the levels ⬇️
The Six Levels of Digital Transformation
Level 1: Systems of Record
From Paper to Digital
Data lives in databases and ERPs, not filing cabinets.
What's Usually Found Here:
Mature Key Operational Systems (ERPs) > Data Capture > Data Storage > Data Retrieval > Data Management > Data Governance
Reality Check:
Still struggling with data quality? Different departments have different numbers for the same thing? Constantly reconciling data?
Stop reading AI white papers. Fix your foundation first.
Level 2: Systems of Engagement
Where Digital Meets Human
Customers, employees, and partners interact with your systems. CRMs, apps, portals, employee interfaces.
What's Usually Found Here:
Mature Engagement Systems (CRMs, HRIS) > Customer Touchpoints > Employee Interfaces > Partner Ecosystems > Omni-channel
Critical Equation:
No engagement = no data. No data = no AI.
Your AI is only as good as the behavioural data it learns from.
Level 3: Systems of Insight
Making Sense of Data
You answer "What happened?" and "Why did it happen?" with analytics. Dashboards are trusted and drive decisions.
What's Usually Found Here:
Mature Data Warehouse or Lakehouse > Descriptive Analytics > Diagnostic Analytics > Basic Predictive Analytics > Data Visualisation
The Test:
If your analytics are questioned, ignored, or recreated in Excel by sceptical business users, you're not at Level 3.
Level 4: Systems of Intelligence
Where Everyone Thinks "AI Transformation" Starts
ML models predict, recommend, and optimise. NLP understands intent. Computer vision automates inspection.
What's Usually Found Here:
Advanced Predictive Analytics > Prescriptive Analytics > Pattern Recognition > NLP > Computer Vision > ML Models > RPA
Critical Insight:
You can't skip here from Level 1.
Without clean data, rich engagement, and trusted analytics, your AI models produce confident predictions based on garbage data and recommendations nobody trusts.
You're solid at Level 4 when:
ML models in production improve outcomes
MLOps practices established
Models monitored and maintained
Data scientists work with domain experts
Level 5: Systems of Action
When Intelligence Becomes Operational
AI doesn't just suggest, it executes. Dynamic pricing adjusts in real-time. Intelligent routing optimises automatically. Personalisation happens without human intervention.
What's Usually Found Here:
Automated Decisioning > Dynamic Optimisation > Intelligent Process Automation (RPA + AI) > Personalisation > Reinforcement Learning
Reality:
This requires trust in your systems, clear guardrails, robust monitoring, and willingness to let AI make decisions that impact customers and operations.
You're solid at Level 5 when:
AI makes thousands of micro-decisions daily
Automated actions improve over time
Clear escalation paths exist
Measurable business impact
Level 6: Systems of Adaptation
Self-Evolving Systems
Systems learn from outcomes and evolve themselves. Models retrain automatically. Interfaces adapt to user behaviour. Processes optimise based on feedback.
What's Usually Found Here:
Continuous Learning > Self-Optimisation > Adaptive Interfaces > Closed-Loop Feedback > Meta-Learning
Hard Truth:
Most organisations will never get here. Level 6 requires sustained investment, exceptional talent, and capabilities that take years to develop. Google, Amazon, Netflix operate here. Focus on mastering Levels 3-5.
Where Are You Actually At?
Level 1: Data quality dominates meetings. Different departments have different numbers. Teams spend more time fixing data than analysing it.
Level 2: Digital channels exist but aren't integrated. User adoption is a challenge. Customer journeys are disconnected.
Level 3: You have dashboards but people create their own reports. Analytics exist but don't drive decisions.
Level 4: ML models exist but aren't in production. AI initiatives have unclear ROI. Data scientists work in isolation.
Level 5: AI suggests actions but humans approve most. Automation is limited. You're cautious about autonomous decisions.
The Path Forward
1. Assess Honestly Where are you actually at? Not where you want to be.
2. Solidify Your Current Level Shaky foundations doom everything built on top.
3. Focus on the Next Level If you're at Level 2, work on Level 3, not Level 5.
4. Integrate, Don't Separate AI transformation is part of digital transformation, not separate.
5. Track What Matters Each level has specific indicators. Measure them.
Before You Launch Another "AI Transformation"
Ask three questions:
Which level are we at today?
What needs to be solid before we progress?
How does this build on our existing digital foundation?
Can't answer these clearly? You're about to waste money and time on an initiative that will fail.
AI isn't a separate destination. It's a natural evolution that only works when you've built the proper foundation.
The question isn't whether to pursue AI. The question is: are you ready?
👉 Follow me for more on building, pricing, and strategising digital products, transformation, and strategy: https://www.linkedin.com/in/arsedian
