You registered on DND, yet your phone still buzzes with loan, insurance and “congratulations you won” calls every single day. From late March 2026 you no longer have to do anything to fight back. By a TRAI Direction dated 27 February 2026, every telecom operator must now run AI/ML systems that detect spam callers on their own and act against them even if no customer ever complained, and must warn other operators about a suspect number within two hours of flagging it.
Until now, action against a spam caller usually started only after a consumer complained on 1909 or the DND app. TRAI found that this complaint-driven model was a weak deterrent, and that roughly 85 percent of spam complaints were against unregistered telemarketers. The new Direction shifts the burden to the operators: their machines must spot the spammer first, and a single flagged number can trigger network-wide action across all telcos.
Here is what the 27 February 2026 Direction makes mandatory, in plain terms:
The Direction was issued under section 13, read with sub-clauses (i) and (v) of clause (b) of sub-section (1) of section 11, of the TRAI Act, 1997. All operators were directed to comply within thirty days from the date of issue, that is by late March 2026.
The new AI system works in the background, but your existing rights to complain are untouched. If a spam call or SMS slips through, you should still report it so the system gets more signals:
Because the system now acts on calling patterns rather than complaints, an ordinary heavy caller, a small business owner, or someone running an OTP-style alert service could in theory be flagged. The Direction builds in a warning step before any harsh action, so do not panic:
As Kashvi Pathak notes, the safeguard for honest users is the mandatory warning notice and the graded KYC ladder, so anyone wrongly flagged should engage with the notice immediately rather than ignore it.
No. From late March 2026 your telecom operator must use its AI/ML system to detect and act against spam callers on its own, even if no customer complained. Your complaints on 1909 and the DND app still help, but they are no longer the only trigger.
The Direction is dated 27 February 2026 and gave all access providers thirty days from the date of issue to comply. That means the system was due to be operational by around late March 2026.
The Direction lists behavioural signatures, not message content, as the basis for flagging. The system analyses patterns such as call and message volume, velocity, diversity, duration and temporal patterns to spot bulk spam behaviour.
Very fast. When a number is flagged as a “Suspected UCC CLI”, the terminating operator must share it with the spammer's own operator over the common DLT platform immediately, and in any case within two hours of flagging.
No instant block. You first get a notification that your number has been flagged. A first flag triggers KYC re-verification, not disconnection. Disconnection and one-year blacklisting apply only to senders found actually sending unsolicited commercial communication.
Use the phone number and email given in the flagging notice to seek clarification and confirm your identity, complete any KYC re-verification within the stated window, and keep the notice and your reply as proof. If service is wrongly cut, escalate with a written complaint or an RTI.
It was issued under section 13, read with sub-clauses (i) and (v) of clause (b) of sub-section (1) of section 11, of the TRAI Act, 1997, and builds on the Telecom Commercial Communications Customer Preference Regulations, 2018.
No. The existing DND framework, the 1909 reporting channel and DLT registration of senders continue. The new Direction adds proactive, AI-driven detection on top of them so action does not depend on your complaint alone.