What is a TRAI compliant voicebot?

A TRAI compliant voicebot is an automated AI calling solution that enforces National Customer Preference Register (NCPR/DND) registry screening, restricts calls strictly to permitted hours (9 AM–9 PM for sales, 8 AM–7 PM for collections), dials from registered Indian telecom CLI numbers, and maintains auditable call records.

Does registering on DND block all calls, including from my bank or lender?

No. DND registration blocks unsolicited promotional/marketing calls (UCC), not transactional communication. EMI reminders, delivery updates, and calls related to an existing service relationship are generally exempt, provided they stay transactional in content.

What are the permitted calling hours for AI voice campaigns in India?

Promotional/commercial calls are restricted to 9:00 AM–9:00 PM under TRAI rules. RBI-regulated entities like NBFCs and lenders typically apply a stricter 8:00 AM–7:00 PM window for collections and recovery calls.

Can a customer opt out of future AI calls during the call itself?

Yes, and a compliant system should honor it immediately. When a customer says something like "do not call me again," the AI agent should acknowledge it, end the call politely, and add the number to an internal do-not-call list so future campaigns automatically skip it.

How do we know whether our campaign is promotional or transactional?

Communication with your own existing customer about their own account, order or appointment is generally transactional; outreach offering something to someone who is not already a customer for it is generally promotional. If a service call also carries a pitch, that can change the classification. When uncertain, treat it as promotional and screen it — and take the boundary cases to your own compliance team.

Does using AI rather than human callers change what is permitted?

No. The rules govern the communication, not who or what places it. In practice a scripted agent makes compliance easier to demonstrate, because mandated disclosures are delivered identically on every call and every call is recorded — which is precisely what human floors most often fail on.

TRAI Compliant Voicebot Guide: Is AI Calling Legal in India?

As artificial intelligence shifts from a support mechanism to active outbound telecalling, businesses in India are eagerly adopting AI voice agents. However, a major concern arises for compliance and legal teams: Is outbound AI calling legal in India? and what are the TRAI compliance guidelines for automated voice bots?

Direct Answer: Is AI Calling Legal in India?

Yes, AI calling is completely legal in India, provided campaigns adhere to the regulations established by the Telecom Regulatory Authority of India (TRAI) and the National Customer Preference Register (NCPR). The legal compliance of your campaign depends entirely on the type of call (transactional vs. promotional), DND scrubbing status, and permitted calling hours.

Key TRAI Compliance Checklist for AI Calling

To run legal, compliant outbound voice campaigns, your business must satisfy these four core regulatory rules:

1. DND (NCPR) Scrubbing: Under TRAI rules, you cannot place unsolicited commercial communications (UCC) to numbers registered on the Do Not Disturb (DND) list. All cold-outreach campaigns must scrub contact lists against the NCPR database before dialing. Vistara AI supports direct integration with list-scrubbing tools.
2. Permitted Calling Windows: Promotional campaigns are strictly restricted to the hours of 9:00 AM to 9:00 PM. Placing automated marketing calls outside of this window is a direct regulatory violation.
3. Transactional & Opt-In Calls: If a customer submits a web form requesting a callback, or if the call relates to an active transaction (like an EMI reminder, service appointment confirmation, or delivery alert), the DND restriction does not apply. These are classified as transactional or explicit-consent communications.
4. Verified Caller ID (CLI): You must dial from registered, legitimate telecom CLI numbers. Using international virtual numbers to mask caller identity is illegal and leads to carrier blocking.

Best Practices for Compliant Voice AI Campaigns

To avoid compliance violations and ensure high customer trust, always implement these practices:

Immediate Opt-Out: Allow the customer to say "Do not call me again". The AI agent should recognize this intent, apologize, terminate the call, and automatically add the number to an internal blacklist.
Identify the Entity: The AI agent should state the business name in the first two sentences (e.g., "Namaste, main Vistara Motors se bol rahi hoon...") so there is zero ambiguity.
Keep Auditable Logs: Maintain detailed transcripts and call recordings. This serves as proof of compliance and consent in case of customer disputes.

DND (NCPR) Registration: What It Actually Blocks

The National Customer Preference Register (NCPR), commonly called the DND registry, lets subscribers opt out of receiving Unsolicited Commercial Communication (UCC) — promotional calls and SMS. Registering on DND does not block every call from every business. It specifically blocks unsolicited promotional/marketing outreach. Calls tied to an existing service relationship — an EMI due date, a delivery update, a support callback the customer requested — fall under transactional communication and are exempt, provided the content of the call stays transactional and doesn't pivot into a sales pitch.

What Happens If You Call a DND Number by Mistake?

Telecom operators can levy financial penalties on the originating entity for UCC violations, and repeat violations can lead to blacklisting of the calling number or entity across the telecom network. This is why automated DND scrubbing — checking every number against the current NCPR database immediately before dialing, not once when the list was first uploaded — is non-negotiable for any serious outbound calling program, human or AI.

How Vistara AI Enforces DND-Compliant AI Calling

DND compliant AI calling is built into the platform rather than left to the customer to configure correctly:

Automatic NCPR scrubbing against the live registry before every promotional campaign dial, with un-scrubbable numbers automatically excluded.
Calling-hour enforcement so promotional campaigns can only be scheduled within the 9 AM–9 PM window, with transactional/service campaigns following the stricter RBI 8 AM–7 PM window where applicable (e.g. lending).
Consent and call-type tagging so transactional calls (EMI reminders, delivery confirmations, requested callbacks) are correctly classified and exempted, while everything else is scrubbed.
Full audit trail — every call is logged with timestamp, DND-check result, and transcript, giving compliance teams a ready answer if a regulator or customer disputes a call.

Promotional vs Transactional: The Classification That Decides Everything

The single most consequential compliance decision is which category a campaign falls into, because it determines whether DND and NCPR screening applies at all. Getting it wrong is a regulatory exposure rather than a technicality, and it is a judgment your business has to make rather than one a platform can make for you.

Generally transactional or service: Communication with your own existing customer about their own account, order or appointment — a delivery confirmation, an appointment reminder, a payment due notice on an existing loan.
Generally promotional: Outreach offering a product or service to someone who is not already a customer for it, including cross-sell to existing customers of a different product.
The grey area: A renewal call that also pitches an upgrade, or a service call that introduces a new offer. Attaching a promotional element to a service call can change its classification.
The safe default: If you are uncertain, treat the campaign as promotional and screen it. The cost of unnecessary screening is a slightly smaller list; the cost of missing required screening is considerably higher.

Worth stating plainly: this page describes how the platform is configured and what businesses commonly do. It is not legal advice, and for anything close to the boundary — particularly in lending, insurance and collections — the classification decision belongs with your own compliance or legal team.

The Controls to Configure Before Your First Campaign

Compliance for AI calling is primarily a configuration exercise. These are the settings that carry most of the weight, and they should be in place before the first dial rather than added after a complaint.

Registry screening: Promotional campaigns filtered against DND and NCPR before dialling, not after.
Calling windows: Hard-blocked outside permitted hours, with queued calls rolling forward to the next window rather than firing at night.
Identification: The agent states which business it is calling on behalf of, at the start of every call.
Global suppression: An opt-out anywhere suppresses everywhere and writes back to your CRM — never held only in the campaign that recorded it.
Frequency caps: Limits per number per day and per campaign, so a retry loop plus a duplicated list cannot become harassment.
Retention and retrieval: Recordings kept long enough to produce a specific disputed call during a grievance or audit — and the retrieval path tested before you need it.

Additional Rules for Lending and Collections

For NBFCs, banks and fintechs, RBI expectations on recovery agent conduct apply on top of TRAI's communication rules, and they apply regardless of whether the caller is a person or a system. Three requirements are worth singling out because they are where human floors most often fail and where a scripted agent has a genuine structural advantage.

First, tone: no threats, no intimidation, and no implied consequences beyond factual contractual terms. A fixed script delivers this identically on every call and every call is recorded, which is far more auditable than a variable human floor. Second, third-party non-disclosure: if someone other than the borrower answers, the debt must not be disclosed — configure the agent to ask for the borrower and end politely if they are unavailable. Third, dispute and hardship routing: a borrower disputing an amount or reporting genuine difficulty must reach a human, never a scripted rebuttal.

Conclusion

Automating your telecalling campaigns with AI voice agents is a powerful way to scale operations, but regulatory compliance must never be an afterthought. By utilizing a secure, DND-compliant AI calling platform like Vistara AI and adhering strictly to TRAI calling hours and DND filtering, you can unlock the benefits of generative voice agents legally and ethically in India.

Is AI Calling Legal in India? TRAI DND Compliance Guide for Voice Bots (2026)