Sebi's department heads to use Gen AI to help with compliance monitoring

The regulator has overhauled their approach for ease of doing business.

The market regulator will be use generative artificial intelligence (GenAI) particularly to scan documents for check for compliance requirement .

At a conference organised by Association of Investment Bankers of India (AIBI), Securities and Exchange Board of India (Sebi) chairperson Madhabi Puri Buch talked about the incident that led to the decision.

Recently, during an internal review , a younger officer made a short presentation to Buch. It was the first real application of GenAI at the regulator.

"It amazed me," said Buch. "He had developed a complete mechanism using ChatGPT to basically read an annual report of a REIT (real estate invesmten trust) and create from it a complete internal Sebi memo across 12-30 regulations to say whether the REIT has complied with the regulations (shown in a table) and the page of the document that carried the evidence of compliance," she added.

The officer said that he took two weeks to make it. Buch was so impressed that she immediately set up a task force.

Also read: Sebi to soon propose norms for funds, key personnel to stop misuse of AIF structures

She asked the head of every operation to see how their department could do something similar, even to check the compliance level of initial public offering (IPO) documents. The approach would now be widely used at the regulator.

In the months to come, 80 percent of the officer's work would be done through automation and the officer would have to apply their mind to more substantive issues on the documents, she said.

Also read: Retail investors shouldn’t rush into IPOs, wait for prices to stabilise, says SEBI chief Buch

In a later session, Sebi's whole-time member (WTM) Anant Narayan said that GenAI would also be used to scan the private placement memorandum (PPM) documents.

The regulator has overhauled its approach to ease of doing business. Buch has spoken at various occasions on how Sebi reduced pendency of applications by tracking them from the date of receipt instead of the date of last information received.

When pendency is tracked by date of last information received, it opens a loophole for officials. As their deadline nears, they ask for one more bit of information.

This escape hatch has now been closed at Sebi, through an internal process.