Digital Transformation Isn't Gone—It's Awaiting Its Acceptance Slip

The article opens with a provocative observation: digital transformation hasn't disappeared; it has simply reached the stage where organizations must produce an acceptance slip (验收单) — concrete, verifiable proof of value.

Digital transformation isn't suddenly unmentioned; it's time to hand in the acceptance slip.

Despite the buzzword's decline, investment continues. Citing the China Academy of Information and Communications Technology (CAICT) 2025 report, China's digital economy reached 59.2 trillion yuan in 2024, accounting for 43.8% of GDP and growing 5.5 percentage points faster than nominal GDP. Cloud migration, process redesign, data integration, and dashboards are still being built, and terms like smart manufacturing, AI+, and data elements are advancing. Yet the phrase "digital transformation" itself has become awkward because stakeholders now ask: Which process? Which action? How much cost reduction? Who confirms value? Who owns the outcome post-launch?

The core issue: the label was convenient because it allowed talking direction first, results later. Now organizations demand results.

I. Fewer Words Doesn't Mean the Work Stopped

Common explanations — AI stole the spotlight, the concept is stale, vendors changed terminology — are true but insufficient. A term's exit doesn't equal failure of the underlying work. No one says "electrification transformation" anymore, yet every company uses electricity. Similarly, "informatization" and "mobile internet transformation" faded because those capabilities became default infrastructure. Successful digital transformation likewise stops being a slogan and becomes measurable operating metrics: customer-service deflection rate, inventory turnover, claims processing speed, R&D cycle time, risk-control loss reduction, operating cash flow.

Ping An Insurance exemplifies this shift. It no longer touts "digital transformation"; it reports numbers for 2024:

AI agent service volume ~1.84 billion interactions, covering 80% of customer service.

93% of life-insurance policies achieve second-level underwriting.

Property-insurance anti-fraud system intercepted nearly 12 billion yuan in losses in one year.

These are math problems, not language exercises. The silence around the term can mean either "done and embedded" or "never delivered, afraid to mention." The only way to distinguish is whether an acceptance slip can be produced.

II. The Real Problem: No Acceptance Slip

The term "digital transformation" was useful because its breadth could encompass any tech project — cloud, middle platforms, apps, process digitization, data governance, dashboards. Early on, "having a system" was progress. But as systems proliferate, organizations naturally ask: Did processes shorten? Are responsibilities clearer? Are decisions faster? Is the frontline lighter? Can business results be confirmed?

These questions expose the term's weakness: it sounds like a direction, not a result. Direction can be perpetually correct; results must pass acceptance. The problem isn't that digital transformation can't be accepted — metrics like cycle time, cost, automation rate, customer experience, ROI are all verifiable. The problem is that many enterprises never defined "what done looks like" at the outset. Without a falsifiable success definition, projects can't be judged successful (no clear target) nor failed (always "on the road"). They are quietly shelved, rebranded, or silenced.

In listed companies, financial institutions, and public sectors, vague transformation slogans increasingly fail budget and audit scrutiny. They must answer: What was done? Was it worth it? Is it controllable? The first two determine budget; the third determines risk. "Still transforming" used to be a free pass for continued funding; now it's losing credibility.

Even when metrics exist, admitting shortfall is hard. The biggest loophole isn't missing systems — it's the absence of an acceptance slip everyone recognizes. Not that no one wrote targets, but that no one is willing to speak truth to those targets after the fact.

III. Why "Installing Systems" Masqueraded as "Transformation"

The industry mantra "digitalization is the means, transformation is the end" is correct but vacuous. The hard part is writing down "what transformation actually looks like." Many projects succeed in project-management terms: requirements signed, system live, processes running, data ingested, docs archived, PPTs presented. But that only proves project completion, not organizational transformation.

Project acceptance asks: Is the system built? Are functions there? Can pages be clicked? Business acceptance asks: Did business actions change? Did processes shorten? Did costs drop? Is the frontline lighter? Too many projects substituted the former for the latter, leading to familiar dysfunctions: system live but business still coordinates offline; process closed-loop but nodes still rely on human chasing; data connected but definitions still aligned in pre-meetings; dashboards published but decisions still gut-feel.

This isn't laziness — it's a misaligned acceptance object. You accepted whether the system runs, not whether the business changed. Digital teams build systems but don't own process definitions, data standards, dashboard-driven decisions, or operational accountability. They carry the pressure to prove transformation value without the authority to define transformation outcomes.

Hence the easiest win is go-live; the hardest is post-go-live. Six months later, promised cost reductions haven't materialized. Retrospectives become blame-shifting circles: business says IT built it; IT says business uses it; data says I only provide data; AI says I provide intelligence; the lead says let's sync again. No one says "this didn't work." Because from day one, the acceptance slip never specified who signs it.

It Doesn't Necessarily Reduce Actions, But It Increases Visibility

Managers love digitalization not because it cuts actions but because it makes things visible: processes, nodes, durations, owners, traces, rankings. Visibility has value — without it management loses control. But if a digital system only adds visibility without cutting wasteful actions, the organization gets heavier. Frontline staff experience: one form becomes three systems; one signature becomes online flow plus offline coordination; one Excel becomes requirement-ticket-develop-deploy; one conversation becomes ticket-screenshot-reason-wait-for-closure. Systems become new management layers; tickets become blame-shifting tools; traces become self-protection; reports become decoration. Digitalization turns from efficiency tool into organizational burden.

Good digitalization reduces actions; bad digitalization adds traces.

Systems change how an action is recorded; only mechanism change alters how the action occurs. McKinsey data: ~90% of large enterprises launched digital and AI transformations, yet average value capture is only ~30% of expectations. Not because they didn't start, but because no line was drawn between "launched" and "done."

IV. Real "Transformation Done" Means Mechanisms on Track

Huawei distinguishes: informatization records process nodes and results into systems; digitalization puts operations, decisions, management, and command onto digital rails. This yardstick is a mirror, not a club — most firms lack Huawei's organizational muscle, process heritage, and sustained R&D investment. Copying Huawei's complex methodology is pointless; its methods are largely post-hoc codification of pre-existing strengths.

The practical takeaway: before starting, write the value clearly. If you can't, don't package it as transformation. This mirror reminds us:

System live ≠ mechanism live.

Data connected ≠ accountability connected.

Pretty UI ≠ lighter organization.

True digitalization asks four questions:

Business objects: Can the system uniquely identify and continuously track them, or just a field?

Business process: Does the system follow the whole process with feedback, or only capture the final result?

Business rules: Are rules encoded as system-enforced constraints, or still in documents relying on human judgment?

Business results: Does someone in the organization own the outcome? Who's responsible for slowness, errors, post-launch operations?

System digitalization solves "where the process runs." Mechanism digitalization solves "why it runs this way, who makes it run, who's accountable when it's wrong." The latter is far harder because systems are easy to change, but authority/accountability, KPIs, departmental interests, and responsibility boundaries are not.

Without mechanism digitalization, no matter how many systems, you've only wrapped the old organization in a new interface.

V. Harder Than Writing Standards: Who Will Speak Truth to the Standard

The mechanism layer is difficult not technically but because the mechanism lacks a position empowered to tell the truth. This gap hides at both ends.

Initiation End: Rewards Vague, Punishes Concrete

Concrete targets ("cut headcount 30% in one year") make approvers hesitate and writers nervous. Vague targets ("comprehensively enhance digital operation level, empower high-quality business development") sail through. So the real acceptance standard never hits paper from day one. The mechanism rewards vagueness and penalizes specificity.

Closure End: Hard to Fail Yourself

At results time, negative feedback rarely reaches the table. The same people who championed, built, and now review the project struggle to declare their own project a failure. Projects aren't killed; they're "continuously advanced." They aren't reviewed; they're "iteratively upgraded." Cycles repeat: rename the initiative, rotate the lead, relabel the budget, change the reporting vocabulary — previous round never closed, next round kicks off. Changing the label equals a fresh spreadsheet, a new game. "Digitalization" becomes "data intelligence" becomes "AI" — new budget secured, previous unmet promises quietly buried. The mechanism makes "keep investing" the safest option and "stop" the hardest.

Can the Gap Be Closed? Change Rules, Not People

Honestly, it's hard. At least three rule changes are needed:

At initiation, write "done" as a falsifiable statement — not "improve management efficiency" but "reduce X waiting time, eliminate Y forms, compress Z approvals, improve metric W." If you can't say it, don't open PowerPoint.

Don't let the driver self-certify success. Value realization needs a relatively independent validator. At minimum, the scorer shouldn't be the same person who approved the project.

Next tranche of money depends on last tranche's results. No closure on previous round, no green light for next. Want new money? First acknowledge old outcomes.

The logic is simple; execution is hard because someone must genuinely enforce it, and workarounds are endless. Even if enforced, it only mitigates, never eliminates. Larger projects involve more departments, diffuse goals, scattered accountability — finding someone who knows the business, is independent, and dares to speak truth becomes nearly impossible. So the gap isn't about unwillingness; the mechanism simply lacks a role that can both clarify and courageously declare results.

No system can plug this gap. Slogans like "value retrospectives" can't either. What's truly missing is deciding on day one: who will sign the acceptance slip — not for "system live," but for "did the promised business change actually happen?" If that step isn't fixed, even a beautiful mechanism is just another unsigned acceptance slip.

VI. AI Isn't a Savior; It's the Hand Flipping Old Accounts onto the Table

This looks like a deadlock: no one writes concrete targets, no one signs, old rounds never close — so we just rename and restart. The new label is AI.

Will AI repeat the pattern? Half yes, half no. AI makes digitalization more critical, but it doesn't want the "digital transformation" narrative. It needs a usable foundation: clean data, connected systems, standardized processes, clear permissions, callable knowledge, traceable accountability chains. Just as "why did big data suddenly go quiet?" revealed AI exposing data's old debts, "why is digital transformation fading?" reveals AI exposing mechanism debts.

Processes not standardized → agents don't know which rule to execute.

Permissions undefined → agent doesn't know who authorizes, who backs up.

Knowledge not codified → agent stitches answers from fragmented docs.

No exception handling → agent crosses boundary, no one dares use it.

No feedback loop → agent becomes less trustworthy with use.

This proves the earlier digital foundations were real but uneven. AI consumes real foundations; it chokes on the unstandardized, unowned portions.

At the "fully embrace AI" slogan layer, it becomes another big bucket. AI-ready, AI-native, intelligent business reconstruction can become new safe rhetoric. Gartner already places generative AI in the "trough of disillusionment." Without clean data, smooth processes, clear permissions, and defined accountability, AI is just a better chat demo.

But AI differs in one crucial way: it pulls acceptance down to the concrete task site. Can this customer-service assistant resolve issues? Can this text-to-SQL app answer with correct definitions? Can this quality-inspection model improve accuracy? Can this agent run the end-to-end process? These questions are far more verifiable than "did we complete digital transformation?" Previously you could say "still on the road." With AI use cases, it either runs or it doesn't. Because this ruler is harder, transformations that only swapped labels without settling old debts will increasingly fail to hide.

A new label can cover things for a while. It can't cover the result that AI produces the moment it's put to work.

VII. Stop Asking If There's Transformation; Ask If You Can Produce the Acceptance Slip

The issue is simple. Digital transformation isn't less important; it's too important to remain a slogan, a project bundle, or a PPT bullet. The first half was a language exercise: built platform capabilities, connected processes, empowered management, supported strategy. None wrong, but no longer enough. The second half demands acceptance slips.

No process account → actions didn't change.

No value account → results unconfirmed.

No accountability account → post-launch ownership missing.

Before approving a project, don't ask what tech it uses or whether it counts as "transformation." Ask five harder questions:

Which business action does it change?

Which real burden does it reduce?

Which operating result does it improve?

Who confirms that result?

If it fails, who has the authority to say it failed?

If these five can't be answered, the hotter the new buzzword, the safer the old problems remain. The lesson isn't to stop building systems or chasing AI. It's to stop paying for initiatives that can't produce an acceptance slip and have no one to sign it.

Systems can go live; processes may not be restructured. Data can connect; accountability may not. Interfaces can look pretty; the organization may not get lighter. AI can be plugged in; the business may not truly change.

Digital transformation didn't suddenly go quiet. It wasn't debunked. It's acceptance-slip time.

Done right: accounts can be rendered, so no need to keep mentioning it. Not done: accounts can't be rendered, so afraid to mention it again.

The real task isn't to bury the term but to distinguish which silence you're in. So this time, don't just write down "what done looks like." On the day you write it, also decide: who will sign that acceptance slip. Otherwise, the next buzzword is already coming. And when it arrives, if still no one will sign, someone will probably write: "Artificial Intelligence+, Why Is No One Talking About It Anymore?"