AI Is Rewriting Delhi's Business Playbook — But the Risks Are Real and the Ethics Unresolved

More than 4,200 small and medium enterprises registered in Delhi's NCR corridor have adopted some form of AI-assisted tool in the past eighteen months, according to a June 2026 survey by the Confederation of Indian Industry's Delhi chapter. The number sounds like a victory lap. Speak to the business owners themselves, and the picture gets complicated fast.

The context matters. India's Union Budget for 2026-27, announced in February, earmarked ₹2,400 crore specifically for AI adoption incentives targeting MSMEs. Delhi, with its dense concentration of tech talent graduating out of IIT Delhi in Hauz Khas and Netaji Subhas University of Technology in Dwarka, was always going to be ground zero for that money hitting the street. But cash and capability are different things, and the gap between them is where the trouble lives.

Promise on Paper, Pressure on the Ground

Walk through Nehru Place, still the nerve centre of Delhi's electronics and software trade, and you hear two conversations running simultaneously. One is about efficiency — retailers using AI inventory tools cutting overstock losses by roughly 18 percent, logistics firms in Okhla Phase II shaving delivery times using route-optimisation models. The other conversation is quieter and considerably more anxious. A textile export house near Lawrence Road, which employs around 340 workers, began piloting an AI-driven quality-inspection system in March. By May, it had laid off 27 floor inspectors. The owner has not spoken publicly. The workers have.

The Gig Workers' Collective of Delhi, which operates out of a shared office in Lajpat Nagar, has documented 140 displacement complaints from workers in manufacturing and back-office roles since January alone. Their concern is not hypothetical. It is ₹18,000-a-month jobs disappearing into algorithms that cost a company a flat ₹6,000-per-month SaaS subscription. The arithmetic is brutal and the policy response has lagged badly.

Ethical risk compounds the economic one. Several Delhi-based fintech lenders operating under RBI's digital lending framework have integrated AI credit-scoring models trained predominantly on urban, English-language data. Consumer advocacy groups at the Delhi Consumer Forum flagged in April that applicants from outer districts — Narela, Bawana, Alipur — were being systematically underscored, effectively denied loans at rates their urban counterparts receive. The models were not designed to discriminate. They did anyway.

Who Watches the Machine?

India's Digital Personal Data Protection Act, which came into force in late 2024, sets baseline rules around data use but contains no specific provisions governing algorithmic accountability or automated decision-making. The government's draft National AI Policy, circulated for comment in March 2026, proposes an AI Governance Board — but that body does not yet exist, and its powers remain undefined in the draft. Businesses adopting AI in Delhi are, for now, largely self-policing.

NASSCOM's Centre of Excellence in Gurugram, which works closely with Delhi-NCR's startup ecosystem, ran its first AI Ethics Workshop in May 2026 with participation from 68 Delhi-based companies. Attendance was voluntary. Compliance with any of the recommendations is also voluntary. That is the regulatory environment businesses are operating inside right now.

The Startup India Hub at Bhikaji Cama Place has begun offering a structured AI readiness audit to registered startups — a twelve-point checklist covering data sourcing, bias testing, and worker-impact assessment. Uptake has been modest: 214 companies completed the audit between April and June, out of an eligible pool of roughly 3,100.

For Delhi's business owners trying to stay competitive, the practical advice is not to wait for regulation to catch up. Audit your training data now. Document which roles are affected before a tool goes live, not after. Run parallel systems for at least one quarter before full deployment. And consult the NASSCOM framework even if it carries no legal weight — courts filling regulatory gaps tend to look for evidence of good faith. The AI opportunity in Delhi is genuine. So is the cost of getting it wrong.