AI is your new teammate. Here's how Operations can lead
India's IT engine is pairing people with AI at scale. Tata Consultancy Services now counts 217,000 employees with higher-order AI skills, up from 180,000 barely a month ago. The message is clear: teams that ship will be human + AI.
A Nasscom-Indeed study says more than a third of Indian IT firms already use AI across 40% of core operations, delivering 25%-35% improvement on key KPIs. By 2027, 97% expect blended teams to be the default, with AI driving most tasks and humans owning judgment, empathy, and hands-on work.
What this means for Ops leaders
- Treat AI as headcount. Design workflows where AI leads routine steps and humans handle edge cases and decisions.
- Redraw SOPs for human-in-the-loop. Define handoffs, escalation paths, and audit trails.
- Shift hiring to "AI natives" while upskilling current teams. Pair junior AI-first talent with experienced domain leads.
- Track value, not hype. Anchor on cycle time, first-pass yield, SLA adherence, error rates, and cost per ticket.
- Stand up governance: data use rules, approvals for prompts/agents, incident reporting, and model-change controls.
The skills gap is real-and it's an Ops problem
Freshers are arriving with AI in their toolkit; nearly a quarter of entry roles now require AI or data skills, up from 5%-10% three years back. Yet India's AI talent supply is about half of current demand, according to NITI Aayog.
Without action, total tech services headcount could slip from 7.5-8 million (2023) to 6 million by 2031. With focused upskilling, it could rise to 10 million. TCS itself reduced headcount from 607,979 (Mar) to 582,163 (Dec) and flagged a "skill mismatch," especially in mid and senior roles.
Forecasts split: some analysts see hiring recovery as AI demand lifts by FY27-FY28, while others argue AI will cut services labor needs sharply. Either way, your hedge is the same-upskill fast and redesign work now.
90-day rollout plan
- Weeks 1-2: Map your top 10 processes by volume and SLA risk. Flag steps fit for AI assist or full automation.
- Weeks 3-4: Pilot 2 processes with clear baselines. Instrument everything (time, quality, rework, exceptions).
- Weeks 5-8: Convert pilots into SOPs. Add human-in-the-loop gates, error taxonomies, and rollback steps.
- Weeks 9-12: Scale to 5-7 processes. Launch a lightweight control board for model updates and policy exceptions.
Team blueprint
Ops leaders can follow the AI Learning Path for Business Unit Managers to structure roles, governance, and expectations for AI-enabled squads.
- AI Product Owner: owns use cases, ROI, and backlog.
- Prompt/Process Engineer: codifies workflows, prompts, guardrails.
- Data Quality Lead: monitors inputs, redacts PII, tracks drift.
- Compliance Partner: reviews risks, approvals, audit logs.
- Squads: 1 AI + 2-5 operators per process lane; rotate a senior reviewer across lanes.
KPIs that matter
- Throughput and cycle time per task
- First-pass yield and rework rate
- SLA hit rate and aged backlog
- Cost per ticket/transaction
- Variance across shifts/teams after AI rollout
- Human touch rate (how often AI needs handoff)
Hiring and upskilling that actually sticks
- Level 1: AI foundations for all operators (prompts, verification, safe data use).
- Level 2: Tool stacks by workflow (code assist, content ops, data prep, customer ops).
- Level 3: Agent orchestration, API handoffs, and failure handling for process engineers.
- Certify and tie to pay bands; publish a skill matrix and staff shifts accordingly.
For operator-facing training and a ready-made curriculum, explore the AI Learning Path for Administrative Assistants. If you need a ready-made curriculum for operators and process leads, also consider this certification track: AI Certification for AI Automation.
Tech and risk checklist
- Model access policy: which models, which tasks, who approves.
- Data handling: PII redaction, retention windows, vendor terms.
- Observability: prompt/version logs, evaluation sets, regression tests.
- Fallbacks: confidence thresholds, human review, safe defaults.
- Change control: track model/SOP changes and their impact on KPIs.
Why act now
AI is already embedded across core operations and pulling KPIs up. Companies that train people to work with it-like TCS-are moving faster than those debating headcount models.
Your edge as an Ops leader is simple: pick the right processes, ship small wins, measure relentlessly, and skill your teams to treat AI as a capable teammate.
Sources: Nasscom-Indeed findings on AI adoption in Indian IT (Nasscom), and policy insights from NITI Aayog.
Your membership also unlocks:
