AI calling compliance in India: TRAI, DND, consent and call timing

Updated 2026-09-19 · 5 min read · By the Convo360 team

Why this needs care before you dial

Outbound calling in India sits under telecom regulation (TRAI) covering unsolicited commercial communication, alongside general consumer-protection and data-privacy expectations that apply to any business collecting or using a customer's phone number. None of this is unique to AI: the same rules that apply to a human calling centre apply to an AI voice agent making the call.

What changes with AI is scale and reach: a bot can dial far more numbers per hour than a team of agents, and it can run unattended overnight if you let it. That makes getting the basics right upfront more important, not less, because a misconfigured campaign can generate a large volume of non-compliant calls before anyone notices.

This section is general information to help you plan and understand the landscape. It isn't a substitute for advicefrom a compliance professional familiar with your business, your call type (transactional vs promotional) and your call volumes. Treat everything below as a starting checklist, not a final answer.

DND, consent and call timing — what to plan for

• DND / NCPR scrubbing: numbers registered under India's National Customer Preference Register generally should not receive promotional calls. Screen your contact list against this before an outbound campaign, and keep transactional/service calls (like an appointment reminder to an existing customer) separate from promotional ones in your thinking, since they're often treated differently.
• Calling hours: stick to reasonable daytime hours rather than early morning or late night, regardless of what a dialler technically allows you to schedule.
• Consent for recording: tell the caller near the start of the call if it's being recorded, and only record where you have a lawful basis to do so for your specific use case.
• Opt-out: give callers a simple way to say 'don't call again' and honour it promptly, both on voice and if you follow up on WhatsApp or SMS from the same contact record.
• Purpose limitation: use the number only for what the customer expects: a customer who gave their numberfor a delivery update may not expect an unrelated marketing call later.

Because rules and enforcement here evolve, and because obligations depend on your specific call type and volume, this list is a starting checklist to bring to a compliance professional, not a complete or final one.

Disclosing that the caller is talking to an AI

Beyond regulation, disclosure is good practice for trust: a short, natural line early in the call, something like 'this is Convo360's AI assistant calling on behalf of [business]' sets expectations and avoids thecaller feeling misled later in the conversation. It also tends to reduce hostile reactions once the caller realises they're not speaking to a person, because they were told upfront rather than figuring it out midway through.

Keep the disclosure brief and then get straight to the reason for the call. A long compliance-sounding preamble often causes the caller to hang up before you've said anything useful, which defeats the purpose of the call in the first place. One short sentence is usually enough: the business name, that it's an AI assistant, and why you're calling.

Record keeping and opt-out on Convo360

Every call on Convo360 is transcribed automatically, and you can trigger post-call actions, an AI summary, a Save CRM lead entry, a team notification, so you have a record of what wassaid and whether consent or an opt-out was voiced during the call. Search transcripts later in Analytics or the CRM if you need to check what a specific caller was told, which is useful both for your own quality checks and for responding to a customer query about a past call.

For WhatsApp and SMS follow-ups from the same contact flow, Convo360 already handles DND / opt-out via STOP keywords in Contacts, so an opt-out on one channel can be reflected in your list rather than living only in someone's memory or a separate spreadsheet nobody checks before the next campaign.

Building disclosure and hours into a call flow

1. Start every outbound voice flow with a play audio or spoken opening line that identifies the caller as an AI assistant and names the business.
2. Use a Condition block or your campaign scheduling settings to restrict outbound calls to your chosen calling-hours window.
3. Add a gather or DTMF menu step early on offering 'press/say to opt out of future calls', and route that to a Save CRM lead or contact-tag update.
4. Route any caller who explicitly objects to a voicemail or graceful call-end block rather than continuing the script regardless of their response.
5. Review Flow Logs and call transcripts periodically to confirm the disclosure and opt-out steps are actually firing as designed, especially after any flow edit.

A simple checklist before you launch a campaign

Before switching on any outbound AI calling campaign, it helps to walk through a short checklist rather than relying on memory: is the contact list scrubbed against DND/NCPR, is the calling-hours window set correctly for your region, does the opening line disclose that it's an AI assistant, is there a working opt-out step, and does someone review a sample of call transcripts afterwards?

Keep this checklist somewhere your team actually looks at before each campaign launch, not just once when the flow was first built. Contact lists, business hours and even your own policies can change between campaigns, and a checklist that only exists in someone's head tends to get skipped under deadline pressure.

None of this replaces a proper review by someone qualified to advise on your specific obligations. Treat this article as the questions to bring to that conversation, not the final word on the answers.

Example conversation

Frequently asked questions

Related reading

Available across India — see WhatsApp & AI calling automation by state and city.