Five Stages of Enterprise Digital Transformation: A Capability Upgrade Roadmap

Informatization: From Manual Records to System Traces

Around 2000, informatization meant moving manual tasks onto computers. Finance used abacuses and handwritten vouchers; sales logged orders in Excel; production workers carried paper work orders across departments for signatures. The core was replacing human operations with computer systems so business processes run online. Typical tools: OA (office automation), ERP (enterprise resource planning), CRM (customer relationship management), and financial software.

Logic: replicate offline processes exactly into online systems.

Example — contract signing: previously three handwritten copies circulated for approval; after ERP, filling one form auto-generates purchase, warehousing, and logistics orders, with approvals in-system.

Efficiency gain: monthly sales consolidation dropped from days to three minutes via ERP export.

Pitfall: many firms install systems but keep old processes. Fields mismatch real business, staff enter dummy data, resulting in messy electronic ledgers worse than paper. That's not informatization — just swapping paper for digital.

Digitalization: From Process Recording to Business Restructuring

By ~2015, firms realized online processes weren't enough; stored data could drive decisions. Digitalization shifts focus from "process completion" to "data usability." Core: make data the basis for decisions, even reshape business with data. Tools: data warehouses (DW), business intelligence (BI), user behavior analysis (UBA), low-code platforms.

Logic shifts from process-centric to data-centric.

Example — marketing channel selection: previously based on experience ("Platform A worked last time"); with BI, calculate ROI per channel — invest 100 on A returns 200, on B returns 80 — so budget shifts to A. Data drives action.

Problem solved: not just efficiency but business innovation.

Pitfall: dashboards become ornaments. One company had 100+ charts; the boss saw sales down 5% but nobody could explain why — product line? region? Dashboards unlinked from actions = not true digitalization.

Intelligentization: From Experience Judgment to Algorithm Assistance

Post-2020, AI rose as firms accumulated massive data (millions of user behaviors, thousands of sensor streams). AI replaces parts of human decision-making, especially complex, repetitive judgments. Examples: factory machine vision detects 0.01 mm defects in 0.1 seconds; bank AI risk models analyze thousands of dimensions instantly for loan approval.

Core: algorithms/models substitute human decisions. Tools: ML platforms, NLP, computer vision (CV), intelligent decision systems (DSS).

Logic upgrades from "data describes past" to "algorithms predict future."

Pitfall: "using AI" ≠ intelligentization. A smart chatbot that only replies "contact human support" for installation questions; an AI quality inspector that flags good products as defective, requiring manual re-check. These are tech showcases, not real help.

Wisdomization: From Point Intelligence to Ecosystem Collaboration

If intelligentization makes one link smart, wisdomization makes the whole system smart. Key: internal data connects, external partner data connects, enabling systems, people, machines to self-coordinate without constant human command.

Break data silos, move from single-link optimization to whole-ecosystem optimization, achieving global auto-collaboration.

Tools: city brains, industrial internet platforms, digital twins. Prime example: Hangzhou's "City Brain" connects 100+ departments (traffic, fire, hospitals, urban management). When a junction congests, system auto-adjusts signal timing, alerts navigation apps to reroute, and opens emergency lanes for ambulances — global self-optimization.

Data Intelligence (数智化): From Tool Assistance to Intelligent Collaboration

Recent buzzword. Not simple digitalization + intelligentization, but deep fusion of data and intelligence, reaching "people use machines, machines help people" — human-machine co-innovation. People offload repetitive work; machines go beyond mechanical execution; together they improve outcomes.

Tools: data middle-platform, AI middle-platform, large language models (LLM) — they bind data and intelligence.

Shift from "process-pushed" to "intelligence-led," radically accelerating innovation speed.

Summary Comparison & Practical Guidance

The article provides a comparison table (image) and advises against blindly chasing stages. No stage is superior; only fit matters. Stage-specific focus:

Informatization (starting): don't rush many systems; first connect core processes.

Digitalization (have systems): prioritize "make data usable" — govern data first, then build "decide by data" habit.

Intelligentization (have models): grab quick-win scenarios; don't chase 100% accuracy or one big model for all.

Wisdomization (have collaboration): identify which data must connect; connect what impacts global efficiency — precision over breadth.

Data Intelligence (deep fusion): remember tech serves business; tech and business must work together.

Many firms waste money chasing stages: buy ERP then demand AI decisions; buy LLMs before cleaning data. Result: stacked systems, unchanged business pain. Core principle: use tech and data to solve real problems. Instead of asking "what stage are we at?", ask "what's our biggest business pain, and which stage's method solves it?" The stage that solves the problem is the right one.