• Member
• May 13, 2026

Together with my job till AI does it apart

A Structural Threshold Foresight on IT Jobs in India (2035)

Framing Premise

Structural Threshold Foresight begins where conventional forecasting ends. Standard foresight extrapolates: it assumes that what is changing will continue to change at roughly its current rate, adjusted for known accelerants. STF refuses this assumption. It asks instead: where are the points of discontinuity? Where does a system not simply change but reorganize — where the rules governing it shift, not merely the values within them?

India's IT workforce is not approaching gradual change. It is approaching a sequence of structural thresholds — four, to be precise — each of which separates two qualitatively different worlds of work. The challenge of STF is not to predict when each threshold will be crossed. It is to identify what the crossing looks like from the inside, so that actors can recognize it before the system moves without them.

The title of this analysis — Together with my job till AI does it apart — is not a lament. It is a structural description. Before each threshold, humans and AI work in tandem. After it, the tandem dissolves. What remains on the far side is a different configuration of work, value, and identity. The question is whether individuals, organizations, and institutions in India will have already relocated before the dissolution occurs.

Threshold Architecture

The four thresholds are sequential but not evenly spaced. Each one creates the conditions for the next. Crossing Threshold 1 does not complete the transformation — it initiates it. The full structural break occurs only when all four have been traversed.

Threshold 1 — Design Cost Collapse

The core dynamic: The marginal cost of producing a software artifact — code, test, documentation, specification — falls effectively to zero.

Pre-threshold state: Software production is labour-intensive at its base. Writing code, generating test cases, maintaining documentation, and performing bug triage all require human time in roughly linear proportion to output volume. This is the economic foundation of India's IT services model: disaggregate work into atomic units, staff each unit with trained engineers at competitive cost, and deliver at scale. Entry-level engineers are the essential substrate. Their value is structural, not exceptional.

Crossing condition — the observable signal: When AI systems generate production-grade code, passing internal review without material human modification, at rates exceeding 60% of routine tasks within a given project type — the economic logic of maintaining large entry-level workforces is broken. This is not a gradual erosion; it is a threshold. Below this rate, human review and correction remain the primary workflow, and AI is an accelerant. Above it, the primary workflow is AI, and human review becomes the supervisory residual. The unit economics of entry-level hiring collapse in the subsequent 18–24 months.

Post-threshold state: Entry-level IT work does not disappear — but it ceases to be a volume category. What survives is entry-level work that cannot yet be codified: ambiguous requirements, novel domain integrations, edge-case handling in high-stakes contexts. The number of such roles is an order of magnitude smaller than the pre-threshold hiring base.

India-specific dynamics: India produces approximately 1.5 million engineering graduates annually, of whom a significant fraction enter IT services in rule-based, process-intensive roles. The GCC (Global Capability Centre) expansion — currently over 1,700 GCCs employing ~1.9 million people — has partially insulated India's mid-to-high skill segment. But the entry-level pipeline, which historically served as the ramp into GCCs, is structurally vulnerable to Threshold 1. The NEP 2020 signal is relevant here: the policy's emphasis on multidisciplinary education may, if implemented with fidelity, produce graduates less optimized for rule-based execution and more capable of operating in ambiguous contexts — precisely the post-threshold requirement. Whether implementation catches up to threshold timing is the critical uncertainty.

Horizon placement: Near to Threshold (2025–2028). Signal already visible in hiring slowdowns at major IT services firms and the internal adoption of AI code-generation tools. The crossing condition is being approached, not yet met uniformly.

Threshold 2 — Batch-Size Independence

The core dynamic: AI systems enable a small team to deliver what previously required a large one, breaking the linear relationship between project scale and workforce size that underpins the IT services revenue model.

Pre-threshold state: IT services pricing and staffing are governed by a logic of proportionality. Larger projects require larger teams. More complex systems require more engineers in maintenance. The revenue model of major Indian IT firms — TCS, Infosys, Wipro, HCL — is built on this proportionality: headcount scales with revenue, and workforce size is both an input and a signal of capacity. Mid-level roles — project managers, system integrators, domain specialists — exist to coordinate this scale. Their value is fundamentally coordinative.

Crossing condition — the observable signal: When AI orchestration systems can manage workflow coordination, dependency tracking, and outcome simulation for projects above a certain complexity threshold — such that a team of 8–12 humans with AI achieves outputs previously requiring 60–80 — the billing model of IT services breaks. This is Threshold 2's crossing signal: not any individual productivity gain, but the point at which project-to-headcount ratios become non-linear system-wide. The first visible sign is not layoffs but the inability to grow revenue through headcount expansion. Revenue grows; headcount plateaus or declines. When this persists for more than two consecutive fiscal years across multiple major firms, the threshold has been crossed.

Post-threshold state: Mid-level roles reorganize around AI supervision, not human coordination. Project managers become AI output validators. System integrators become context-providers who tell AI systems what the client's operational reality actually is — the tacit knowledge that cannot be prompted. Domain specialists persist but shrink: their value lies in the judgment that sits at the edge of what AI can reliably generalize. The industry structure moves from pyramid to barbell: a thin top of strategic and interpretive roles, a residual base of execution oversight, and a compressed middle.

India-specific dynamics: This threshold hits India's IT services model at its structural centre. The pyramid was not merely a workforce shape; it was a career pathway. Engineers entered at the base, moved through middle tiers over 8–12 years, and reached senior positions. Threshold 2 eliminates much of the middle — not instantly, but structurally. The tier-2 city IT expansion (Pune, Hyderabad, Coimbatore, Jaipur) was built on the assumption of continued pyramid staffing. GCCs are more insulated because they were already optimizing for higher-value work. The divergence between GCC-India and services-India will widen sharply past this threshold.

Horizon placement: Threshold (2027–2031). Currently in approach. The signal of revenue-headcount decoupling is beginning to appear in quarterly reports but has not yet persisted systemically.

Threshold 3 — Structural Break

The core dynamic: The operating model of IT firms reorganizes around AI-native workflows — not AI-augmented human workflows. The organizational grammar changes: team structures, decision hierarchies, performance metrics, and billing models all require reconstruction, not adaptation.

Pre-threshold state: Even with significant AI adoption, most firms in the pre-Threshold-3 state are running AI inside human workflows. Humans define the process; AI assists within it. This is augmentation. It improves productivity but does not change the organizational logic. Teams still have the same shape; projects still follow the same phases; deliverables are still scoped and reviewed in the same way. The change is quantitative, not structural.

Crossing condition — the observable signal: Threshold 3 is crossed when a significant cohort of firms reorganizes its operating model — not its toolset — around AI-native workflows. The signal is not AI adoption rates but organizational redesign: firms that flatten decision hierarchies because AI handles information synthesis; that redefine "done" because AI can continuously optimize rather than deliver fixed-scope projects; that shift billing from time-and-materials or fixed-fee to outcome-and-improvement models. When 20–30% of major IT services contracts are structured around AI-native outcomes rather than human-effort proxies, Threshold 3 has been crossed institutionally.

Post-threshold state: The surviving organization looks fundamentally different. It is smaller, faster, and more asymmetric in value distribution. It has fewer layers between problem and solution. It invests heavily in what cannot be automated: client relationship depth, contextual intelligence, ethical oversight, and the capacity to define problems that AI systems can then solve. The role that emerges at the center is not the manager or the architect but the problem framer — the person who can take an ill-defined organizational challenge and translate it into a precisely constrained system that AI can operate on.

India-specific dynamics: India's IT sector faces a structural identity question at this threshold. Its competitive advantage was built on cost, scale, and process maturity. None of these advantages survive Threshold 3 in their current form. Cost arbitrage is irrelevant when the marginal cost of AI output is near zero globally. Scale becomes a liability if it cannot be reorganized. Process maturity helps only if the processes being matured are the right ones. The firms that will survive and grow past Threshold 3 are those that use their existing client relationships and domain depth as the base from which to build AI-native operating models — not those that treat AI as an efficiency layer on legacy structures. This is a strategic choice, and it will differentiate Indian IT firms from each other as sharply as it differentiates India from other geographies.

Horizon placement: Post-Threshold leading edge (2030–2033). Currently visible only in pioneer firms and high-margin GCCs. Will become the dominant mode of high-value IT delivery by early 2030s.

Threshold 4 — Human Judgment as Terminal Value

The core dynamic: The only non-automatable residual is the framing of problems, the arbitration of values, and the interpretation of context at the interface of technology and civilization. Human roles exist not because they are efficient but because they are constitutive — they define what the system is for.

Pre-threshold state: Even past Threshold 3, humans are still deeply involved in executing complex work — designing systems, making architectural decisions, specifying constraints. AI assists and accelerates, but the human is still the primary agent of design.

^{Crossing condition — the observable signal: Threshold 4 is crossed when AI systems can design systems, evaluate their own outputs against multiple value frameworks, and propose architectural alternatives — such that the human's primary contribution is no longer design but purpose-setting. The signal is not AI capability in isolation, but the emergence of a new type of human role: not architect, not manager, not coder — but curator of intelligence ecosystems. When this role begins to appear as a formal job category in organizational charts, and when the people filling it are valued primarily for their judgment about what matters, not their knowledge of how things work, Threshold 4 has been crossed socially even if not uniformly.}

Post-threshold state: Work at the top of the IT stack is no longer about designing solutions. It is about designing the conditions under which AI systems produce solutions aligned with human values, organizational goals, and social constraints. The skills required are not primarily technical: they are philosophical, contextual, and relational. The IT professional of 2035 who survives and thrives is one who can articulate problems with precision, embed ethical constraints structurally, and evaluate AI-generated outputs against criteria that cannot themselves be automated — because the criteria are matters of judgment, not optimization.

India-specific dynamics: This is where India's civilizational depth becomes a structural advantage — if cultivated. India's intellectual traditions in systems thinking, contextual reasoning, and value-embedded decision-making are precisely the capacities that Threshold 4 rewards. But accessing this advantage requires a transformation in educational philosophy: away from certification-oriented curricula and toward the cultivation of judgment, problem framing, and interdisciplinary synthesis. The institutions that make this transition early will produce the professionals who matter most past Threshold 4. Those that do not will produce technically proficient graduates with diminishing structural relevance.

Horizon placement: Post-Threshold (2032–2035 and beyond). Currently visible only in early organizational experiments and academic frameworks. Will define the steady-state of high-value IT work by 2035.

Current Position on the Threshold Map

India's IT sector currently sits between Threshold 1 and Threshold 2, with the crossing of Threshold 1 in progress and the approach of Threshold 2 clearly visible.

Threshold	Status	Key Signal to Watch
T1: Design Cost Collapse	Crossing (2025–2028)	AI-generated code passing review without modification >60% of routine tasks
T2: Batch-Size Independence	Approaching (2027–2031)	Revenue-headcount decoupling persisting across 2+ fiscal years
T3: Structural Break	Distant but visible (2030–2033)	20–30% of major contracts restructured around AI-native outcomes
T4: Human Judgment as Terminal Value	Horizon (2032–2035+)	"Curator of intelligence ecosystems" appearing as formal job category

The most dangerous position for an individual, organization, or institution is to believe that the passage of Threshold 1 is the extent of the disruption — to optimize for the post-T1 world as if it were the final state. Each threshold creates the conditions for the next. The adaptation required at T2 is structurally different from the adaptation required at T1; the adaptation required at T3 makes the previous two look incremental.

The Psychological Threshold

There is a fifth threshold, not in the sequence but beneath all of them. It is the threshold of identity.

For three decades, the IT profession in India was a social contract as much as an economic one. It promised stability, upward mobility, and the dignity of technical expertise. Parents measured progress by placement; cities reorganized around campuses; the word "engineer" carried specific social weight. This contract is not being renegotiated — it is being dissolved.

The psychological threshold is crossed when professionals shift from an identity anchored in what they know to an identity anchored in their capacity to reframe what they know. This is not a comfortable transition. It requires accepting that the skills that brought one to the present will not carry one to the future — and that this is not a failure, but a structural feature of living through a threshold epoch.

Together with my job till AI does it apart is not only an observation about the economy. It is an observation about the self. The phrase holds together the two states — the partnership and the dissolution — in a single breath. The intelligence of that phrase lies in its honesty: the partnership is real, the dissolution is coming, and the distance between them is not a gap to be dreaded but a threshold to be crossed deliberately.

Strategic Imperatives by Actor

For individuals: The adaptation timeline is different at each threshold. At T1, the imperative is to move up the value stack before the crossing, not after. At T2, the imperative is to develop coordinative judgment that AI cannot replicate — the capacity to hold contextual complexity that exceeds what can be prompted. At T3, the imperative is organizational: align with firms that are redesigning, not firms that are optimizing. At T4, the imperative is civilizational: develop the kind of judgment that only a human life, richly lived and rigorously reflected upon, can produce.

For organizations: The strategic error is treating each threshold as a discrete efficiency opportunity rather than a structural reorganization requirement. Firms that use AI to do the same things faster will be overtaken by firms that use AI to do different things entirely. The window for organizational redesign — between T1 and T3 — is approximately five to eight years. It is open now.

For institutions (education, policy): The relevant threshold for educational institutions is T4, because the graduates they produce today will be working in the T4 environment. Curricula designed for T0 — the pre-threshold world — are already obsolescent. The shift required is not from one technical curriculum to another, but from a curriculum of technical knowledge to a curriculum of adaptive judgment. This is the deeper implication of NEP 2020's multidisciplinary emphasis, if that emphasis is taken seriously rather than administered formally.

Closing: What STF Uniquely Reveals

Conventional foresight on AI and IT jobs produces a familiar picture: some jobs will disappear, new jobs will emerge, and the net outcome is uncertain. This is true but not useful. It does not tell actors when to move, what the movement requires, or how to recognize that a threshold is being approached rather than a temporary fluctuation.

Structural Threshold Foresight reveals that the disruption of Indian IT is not one event but four — and that each event is qualitatively different from the ones before and after it. The appropriate response to T1 is not the appropriate response to T3. Acting on T1 logic in a T3 world will produce well-executed irrelevance.

The title's wordplay holds the entire analysis in compressed form. Till — as in duration, and till — as in the moment of severance. Together until the moment apart. The job and the person are in partnership until the threshold is crossed; after it, the partnership reorganizes. STF's contribution is to name the threshold, describe its crossing condition, and give actors enough lead time to arrive on the far side by design rather than by default.

By 2035, India's IT sector will be defined not by those who survived the disruption, but by those who recognized the thresholds early enough to become architects of the post-threshold world. The window is open. The thresholds are approaching in sequence. The only question is whether movement precedes the crossing — or follows it.