Earlier this week, I attended the Infosys Enterprise AI World Tour in Tokyo, listening to CIOs and technology leaders wrestle with a question that has haunted Japanese enterprises for a decade: How do we actually transform with technology?

In 2018, I wrote about Japan’s “digital cliff”: Japan’s Ministry of Economy, Trade, and Industry (METI) warned that legacy systems would become unsustainable by 2025, costing the economy over 2% of GDP annually. I predicted that, when trains started running late, Japanese firms would finally embrace transformation. Seven years later, the trains still run on time, but a different kind of disruption has arrived.

The DX Decade Fell Short

In 2022, we wrote that Japanese firms were losing their appetite for digital transformation (DX). Just 39% were actively engaged in DX, and a stubborn 9% remained completely uninterested. Local Japanese firms in particular seemed frozen, trapped between legacy systems and a risk-averse culture.

Fast-forward to 2025 — the year of the “digital cliff” that METI warned about seven years ago. How has Japan fared? Tomio Kikyoubara of the Nikkei BP Research Institute delivered the verdict: 76.8% of Japanese companies promote DX initiatives, but only about one-third have achieved notable results. Despite all the ministry guidance, industry coalitions, and corporate commitments, DX has fallen short in Japan.

The problem isn’t awareness — it’s the approach and lack of governance. Most firms treated DX as modernization: digitizing existing processes, implementing robotic process automation, and hoping that efficiency gains would compound. But without executive sponsorship and clear accountability, initiatives failed to deliver value — the same processes, the same work, a little bit faster, a little bit cheaper. More than inadequate, this model is now obsolete.

Welcome To AX

Kikyoubara-san gave the shift a name: AI transformation — “AX,” or AI-premised DX. The multinationals at the Infosys event were already demonstrating what this looks like. Roche earned the World Economic Forum Global Lighthouse Award by using digital and AI solutions to cut yield variability by approximately 60%, halving technology transfer time and reducing Scope 1 and 2 emissions by around 31%. AstraZeneca, as another examle, is shifting from campaign-based marketing to AI-orchestrated micropersonalization across channels.

In Japan, the AX narrative lands differently than in the West. There’s no widespread anxiety about AI replacing jobs here: The talent scarcity is too acute. For instance, Compass Group Japan is deploying robotics and AI, not to cut headcount but because it can’t fill roles fast enough in a labor-intensive industry. Japan doesn’t just want AI — it needs it.

What AX Demands

As I’ve argued recently, the real bottleneck isn’t AI capability — it’s organizational reinvention. AstraZeneca’s Kahori Tamura put it best when she said, “AI itself doesn’t drive success. It’s always the people who bring success.” Speaker after speaker returned to the same AI readiness themes. First, many noted the necessity of data readiness — not just structured data — for identifying, organizing, and governing unstructured content such as documents, images, and video. Others focused on governance and trust rails for compliance with the EU AI Act and Japan’s Act on the Protection of Personal Information. Here, Forrester’s AEGIS framework can help. And finally, there was plenty of discussion around new operating models for human-AI collaboration that redesigns roles, not just processes — what we call agent experience.

Roche’s Wilson Wang offered the clearest prescription for escaping proof-of-concept purgatory: “[It’s essential to] have the right sponsorship, the right governance — what to invest, where to invest, and who to invest in.” Roche’s “everyday AI” initiative builds foundational literacy across the organization before moving into higher-investment, business-unit-specific use cases. Wang also stressed establishing master data governance first to avoid automating chaos with different definitions across the business.

Infosys CTO Rafee Tarafdar framed this as “exponential engineering” — using platforms such as Infosys Topaz Fabric to build in days what previously took weeks while enabling integration with existing enterprise investments. For Japanese firms stuck in pilots, this velocity matters.

The Second Awakening

The legendary reliability of Japanese business excellence was built by human coordination at scale. The next era will be built by human-AI collaboration at speed. This time, Japan can’t be late.