TCS CEO Predicts AI Disruption Will Be Deeper and Broader, Tech Industry Transfo

Why Is This ‘Wolf’ Real? How Do Depth and Breadth Define the Next Decade?

Answer Capsule: Because AI, particularly generative AI, touches the “cognitive core” of business operations. Unlike past automation that handled repetitive processes, it directly intervenes in the highest levels of value creation: analysis, creativity, decision-making, and customer interaction. Its breadth is reflected in “pan-industry penetration,” affecting everything from predictive maintenance in manufacturing to risk models in finance.

Whenever a new technological wave arrives, some question whether it is just another overhyped cycle. However, when the CEO of Tata Consultancy Services, one of the world’s largest IT service companies and a witness to decades of technological change, describes AI as “deeper and broader,” we must recognize this not as marketing rhetoric but as an industry forecast based on frontline client needs.

Past technological disruptions, whether the proliferation of personal computers, the rise of the internet, or cloud migration, were essentially iterations of “efficiency tools.” They optimized information flow and reduced computing and storage costs, but most business logic and decision-making processes remained human-led. AI, especially large language models, brings “capability empowerment.” It enables machines to begin understanding, reasoning, and generating outputs that were once exclusive to human domains. This qualitative shift from “tool” to “collaborator” or even “replacer” constitutes the essence of “depth.”

The so-called “breadth” can be observed from penetration rates and impact chains. According to McKinsey’s 2025 research, generative AI technology is expected to affect at least 10% of working hours across over 80% of global occupational categories within the next three years. This means almost no white-collar job can remain entirely untouched. More critically, AI’s impact is chain-like: when software development accelerates due to AI-assisted coding, the innovation cycles of all industries reliant on software updates (i.e., almost all industries) will be compressed, forming an accelerating feedback loop.

timeline
    title Evolution of Technological Disruption Impact
    section PC/Internet Era
        1980s-1990s : Impact Scope: Office Productivity<br>& Information Access
        : Core Change: Process Digitization
    section Cloud/Mobile Era
        2000s-2010s : Impact Scope: IT Infrastructure<br>& Anytime, Anywhere Services
        : Core Change: Service Delivery Model
    section AI Era (Ongoing)
        2020s-2030s : Impact Scope: All-Industry Knowledge Work<br>& Decision-Making Core
        : Core Change: Democratization of Cognitive Abilities<br>& Business Logic Restructuring

The “AI Investment Paradox” for Enterprises: Is Not Spending Now a Death Sentence, While Reckless Spending Is Suicide?

Answer Capsule: The paradox indeed exists. The core solution lies in shifting from “technology experimentation” to “business case-driven” investment. Successful enterprises no longer ask “how to use AI” but “to achieve which specific business objective (e.g., reducing customer churn by 15%), how should we deploy AI.” This requires precise value stream mapping and phased validation.

Krithivasan’s mention that “clients must invest quickly, otherwise they will be at a competitive disadvantage” highlights the widespread anxiety among current business leaders. However, the market is flooded with various solutions from chatbots to complex predictive models. Where should budgets be directed? Many enterprises fall into “shotgun-style” pilot projects, with each department conducting small-scale experiments but lacking coordination, leading to exacerbated data silos, unmeasurable ROI, and ultimately, management losing confidence in AI.

The real strategy lies in “focus” and “integration.” Enterprises need to prioritize identifying key processes with high business value where AI has proven effective. For example, rather than simultaneously investing in AI financial advisors, anti-fraud systems, and credit assessment models, a retail bank could concentrate efforts on conquering one area, establishing a complete closed loop from data pipelines, model training, integration deployment to effectiveness tracking, then standardizing and modularizing successful experiences for horizontal replication.

The table below compares two typical AI investment mindsets and their potential outcomes:

According to Gartner’s forecast, by 2027, over 50% of enterprise AI projects will fail to achieve expected benefits due to lack of clear business cases and integration planning. This underscores the vast execution gap between “having AI” and “using AI correctly.” The value role of system integrators like TCS is shifting from past “outsourced implementation” to “strategic navigation,” helping enterprises chart safe courses on this new continent full of temptations and pitfalls.

Who Are the Winners and Losers? Industrial Tectonic Shifts in the New Competitive Landscape

Answer Capsule: Winners will be “AI-native enterprises” and “traditional giants that rapidly reinvent themselves”; losers will be “bystanders” and “enterprises only superficially digitized.” Additionally, the entire tech value chain will be restructured: cloud giants consolidate the infrastructure layer, model providers compete for the middle layer, while the greatest value capturers may be application-layer companies that deeply integrate AI into vertical industry workflows.

Every deep technological transformation reshapes the industry landscape. AI’s disruptiveness lies in eroding traditional competitive barriers from multiple dimensions simultaneously.

1. Democratization of Knowledge Barriers: In the past, the expertise of top law firms, management consultancies, or financial analysts was a core asset. Now, through professionally trained AI models, small and medium-sized players can also obtain analytical and draft generation capabilities of considerable caliber, compressing excess profits from knowledge monopolies.
2. Redefinition of Economies of Scale: In manufacturing, traditional economies of scale came from hardware production capacity. In the AI era, “data scale” and “model iteration speed” become the new economies of scale. Enterprises with unique, high-quality data flows, even if small in physical scale, can train highly competitive proprietary models.
3. Shift in Ecosystem Control Points: In the PC era, control points were operating systems; in the mobile era, app stores and social platforms. In the AI era, initial control points are considered foundation models. But as models open-source and fine-tuning tools proliferate, application interfaces, workflow integration, and vertical domain data may become more enduring moats.

We can foresee several key competitive axes forming:

mindmap
  root(New AI Competitive Landscape)
    Technology Stack Controllers
      Cloud Infrastructure Giants (AWS, Azure, GCP)
      Foundation Model Providers (OpenAI, Anthropic, Open-Source Community)
      Specialized Chip Designers (NVIDIA, AMD, Custom ASIC Enterprises)
    Vertical Domain Disruptors
      Traditional Industry Leaders with Proprietary Data
      AI-Native Startups (Building AI-First Processes from Scratch)
    Integration & Scalability Experts
      Global System Integrators (TCS, Accenture)
      Service Providers Specializing in AI Deployment & Management

Taking the software development industry as an example, tools like GitHub Copilot have already increased developer productivity by 20-30%. This not only means faster software delivery but may change the game rules: small, agile teams, aided by AI, could challenge projects that previously required massive engineering armies. This directly impacts traditional software outsourcing models centered on man-month billing, forcing service providers like TCS to upgrade their value proposition from “providing manpower” to “providing AI-driven solutions and business outcomes.”

The “Great Skill Migration” in the Labor Market: Are We Replacing Humans or Redefining Work?

Answer Capsule: AI replaces not “jobs” but “tasks” within jobs. This will trigger an unprecedented “great skill migration.” Future work will be a collaborative combination of humans and AI, with human roles increasingly focused on strategy formulation, complex judgment, ethical oversight, creative exploration, and the management and tuning of AI systems.

Discussions about AI and employment often polarize into two extremes: a utopian dream of full automation or a doomsday prophecy of unemployment waves. Krithivasan cites a historical perspective, noting that technological disruption ultimately increases economic activity and employment, provided people are willing to adapt. The connotation of this “adaptation” in the AI era is far more complex than in the past.

Past industrial automation or ITization mainly affected rule-based physical or administrative tasks, allowing the workforce to transition through relatively clear vocational training (e.g., learning to operate new machines or use office software). AI impacts cognitive tasks, requiring the workforce to develop a completely new set of “meta-skills”:

• Prompt Engineering & Iteration: The ability to effectively command AI to produce desired outcomes.
• AI Output Verification & Calibration: Critically assessing the accuracy, bias, and applicability of AI-generated content.
• Human-Machine Workflow Design: Decomposing a job into parts suitable for AI execution and parts requiring human oversight, seamlessly connecting them.
• Data-Driven Decision-Making: Understanding the logic behind AI model recommendations and combining them with human contextual wisdom for final judgment.

The challenge for enterprises and education systems is immense. The World Economic Forum’s “Future of Jobs Report 2025” estimates that by 2027, over 40% of the global workforce’s core skills will need updating. This cannot be solved by a single training course but requires building a culture and infrastructure for continuous learning. Corporate human resource strategies must shift from “position filling” to “skill management,” dynamically tracking skill supply-demand gaps within the organization and bridging them through internal mobility, micro-credentials, and external collaboration.

For individuals, future career security will no longer come from a job title or company seniority but from a portable “skill portfolio” and the adaptability of continuous learning. This is a silent revolution underway, whose social impact in depth and breadth may be no less than AI’s impact on business operations.

Taiwan’s Tech Industry at a Crossroads: The End of OEM Mentality and the Beginning of Value Creation?

Answer Capsule: The hardware manufacturing and efficiency-oriented OEM model that Taiwan’s tech industry excels at may see its value proportion decline in the AI era. The way out lies in upward integration of “intelligent solutions” and downward rooting in “key component innovation.” For example, shifting from “manufacturing servers” to providing “AI model training cluster optimization services,” or from “producing sensors” to developing “edge AI inference modules.”

Taiwan holds a critical position in the global tech supply chain, but in the AI wave centered on software and algorithms, our positioning appears somewhat ambiguous. The past successful model—becoming invisible champions for international brands through卓越 engineering craftsmanship, scale manufacturing, and cost control—is being challenged. Because AI’s value increasingly lies in the integration of software, data, and services, with hardware gradually becoming the “carrier” of intelligence rather than the value core.

This does not mean hardware is unimportant. On the contrary, AI’s insatiable demand for computing power is driving a new wave of hardware innovation, from AI accelerator chips and high-bandwidth memory to advanced cooling solutions. Taiwan’s advantage lies in having a complete industry cluster from semiconductor manufacturing, IC design to electronic assembly. The key is whether we can more tightly bind these hardware advantages with AI’s application value chain?

Opportunities exist in two directions:

1. Become Key Innovators in AI Infrastructure: Not just supplying but participating in defining next-generation AI hardware architectures. For example, designing more efficient specialized chips or system modules for specific AI workloads (e.g., recommendation systems, autonomous driving).
2. Develop AI Integration Capabilities in Vertical Domains: Leveraging Taiwan’s deep experience in manufacturing, healthcare, smart cities, etc., to create “Domain-Specific AI” solutions. Packaging hardware, sensor data, AI models, and industry knowledge into exportable intelligent services.

The table below outlines potential transformation paths and risks for Taiwan’s tech industry in the AI era:

Taiwan’s industry transformation requires not only technological investment but a paradigm shift in thinking: from pursuing “lowest cost” to pursuing “highest value,” from “order-taking production” to “defining specifications,” from “invisible supply chain” to “visible brand ecosystem.” This path is full of challenges, but the industry restructuring brought by AI is precisely the best时机 to break the old格局.