Yes, AI calling is legal in India — but only if it follows the same TRAI DND rules that govern AI calling in India for any automated or promotional outbound call. As of 2026, that means the sender must be registered on the DLT (Distributed Ledger Technology) platform, use approved headers and content templates for promotional messaging, honour DND (Do Not Disturb) preferences, stay inside permitted calling windows, and keep an auditable record of consent. Skip these and you risk header blacklisting, financial disincentives passed on by your telecom operator, and — for regulated lenders — action under the RBI Fair Practice Code and DPDP Act.
This guide breaks down exactly what applies to voice AI outbound in India, what DLT registration for voice AI involves, how consent and scrubbing work, and how a platform like Agni enforces the rules for you instead of leaving compliance to a spreadsheet.
The 2026 Regulatory Stack for AI Calling in India
There is no single "AI calling law." Instead, an AI outbound programme sits under four overlapping frameworks. Understanding which one governs which behaviour is the whole game.
| Framework | Regulator | What it governs |
|---|---|---|
| TCCCPR / DLT + DND | TRAI | Commercial communication: sender registration, headers, templates, consent, DND scrubbing, calling windows |
| DPDP Act, 2023 | MeitY | Personal data: lawful basis, notice, consent, data-principal rights, security |
| Fair Practice Code | RBI | Lender/recovery conduct: no harassment, permitted hours, dignified communication |
| Recorded-voice / robocall norms | TRAI / DoT | Automated and pre-recorded calls, opt-out, identification of the caller |
The critical mental model for 2026: TRAI regulates the channel and consent, DPDP regulates the data, and RBI regulates the conduct. An AI voice agent has to satisfy all three at once.
Do You Need DLT Registration for a Voice AI Campaign?
If your calls are promotional or transactional commercial communication sent at scale — sales pitches, offers, EMI reminders, delivery confirmations, OTP-linked flows — then yes, the principal entity (your business) must register on a DLT platform operated through your telecom operator. DLT registration for voice AI is not different from DLT registration for a human call centre; the obligation attaches to the sender, not to the technology dialling the call.
Practically, DLT onboarding involves four steps:
- Entity registration: register your business as a Principal Entity on the DLT portal of an operator (Jio, Airtel, Vi, BSNL) using PAN, GST, and authorised-signatory details.
- Header registration: register the sender IDs / caller-line identity you will use, mapped to your consent category.
- Content template registration: for messaging tied to campaigns, register the content templates and variables you will use.
- Consent template & category mapping: declare whether traffic is promotional, transactional, service, or explicit-consent-based — this determines the scrubbing and window rules that apply.
Rule of thumb: purely transactional or service calls a customer has an existing relationship for (an EMI reminder to your own borrower, a delivery confirmation) face lighter scrubbing than cold promotional calls to a purchased list. But every commercial programme still needs a registered header and a defensible consent record.
Consent: The Foundation Everything Else Rests On
Under both TRAI's framework and the DPDP Act, consent must be free, specific, informed, and revocable. For AI calling that translates into concrete requirements:
- Purpose-bound: consent captured for "loan servicing" does not authorise cross-sell calls for insurance.
- Auditable: you should be able to show when, how, and for what a number opted in — a timestamp, source, and category.
- Revocable: an opt-out on a call must propagate so the number is not dialled again for that purpose.
- Disclosed: the person should know they are receiving an automated/AI call and who is calling — do not disguise the agent as something it is not.
DND Scrubbing and Calling Windows
Every number carrying a DND preference must be scrubbed against the registry before a promotional call is placed. A fully DND number should not receive promotional traffic at all; a partially DND number receives only its opted-in categories. This scrub is not a one-time exercise — DND status changes, so it must run against a current list at dial time.
Calling windows matter just as much. Promotional and telemarketing calls are restricted to daytime hours and must avoid national do-not-disturb periods. For lenders and recovery, the RBI Fair Practice Code independently confines contact to reasonable hours — typically 8am to 7pm — and prohibits calls that amount to harassment regardless of DLT status. In practice, the safe operating envelope is the intersection of the TRAI window and the RBI window, applied per recipient time zone.
Penalties: What Non-Compliance Actually Costs
The consequences are layered rather than a single fine:
- Header/telemarketer blacklisting: repeated violations lead to your sender ID being suspended or blacklisted by the operator, killing the campaign channel.
- Financial disincentives: under TRAI's regulations, operators levy graded usage/financial disincentives for consent and scrubbing breaches, which are passed through to the offending entity.
- DPDP penalties: data-protection breaches can attract significant monetary penalties adjudicated under the DPDP Act — materially higher than legacy IT-Act fines.
- RBI/regulatory action: for NBFCs and banks, harassment or unfair-practice complaints can trigger supervisory action, not just customer refunds.
- Reputational and legal exposure: recorded, non-consented calls are easy to escalate to a regulator or consumer forum.
How Agni Enforces TRAI/DND Compliance Automatically
Agni, Ravan.ai's India-first voice-AI calling platform, is built so that compliance is enforced by the dialler rather than left to operators to remember. In our deployments this typically means:
- Window enforcement: campaigns run only inside configured calling windows, with per-recipient time-zone awareness and hard blocks outside RBI/TRAI-permitted hours.
- DND and suppression handling: DND lists, internal do-not-call suppression, and per-campaign opt-outs are respected at dial time, and an in-call opt-out immediately suppresses future contact for that purpose.
- Consent and category mapping: traffic is tagged by consent category so promotional flows and transactional/service flows follow their correct rule sets.
- Auditable logs: every call carries a record — number, timestamp, outcome, consent basis, and recording where permitted — so you can answer a regulator with evidence, not assurances.
- RBI-aligned conduct: for collections and EMI use cases, Agni's agents are designed around the Fair Practice Code — no threats, no repeated-call harassment, dignified language in 30+ Indian languages and Hinglish.
Agni is DPDP-aware and built to operate within the RBI Fair Practice Code and TRAI/DND framework, integrating with Indian telephony (Airtel, Twilio, Telnyx, SIP) via no-code flows or REST API. And because compliance shouldn't come at a premium, Agni runs at an all-in rate from ₹2/min — India's lowest — with plans starting at ₹2,999/month and no per-feature stacking.
Compliance is not a feature you switch on at the end. For AI calling in India as of 2026, it is the operating envelope the whole programme has to live inside — registration, consent, scrubbing, windows, and records, on every single call.
Bottom line: register on DLT, capture and honour consent, scrub against DND, stay inside the RBI/TRAI calling window, and keep auditable logs. A platform that enforces these automatically turns compliance from a liability into a competitive advantage.
This article is general information, not legal advice. Confirm your specific obligations with qualified counsel and your telecom operator before launching a campaign.
