Artificial intelligence (AI) is reshaping investment banking workflows. Tasks such as preparing pitch materials and compiling financial profiles, once requiring days of intensive research, can now be completed in a matter of hours. As banks begin testing large language models to draft memorandums and automate due diligence, the industry is confronting structural questions around advisory fees, data confidentiality and regulatory liability.
Mint explains how significant AI’s impact really is on investment banking.
What are the AI use cases in investment banking?
A 2023 Deloitte report noted that global investment banks have invested billions of dollars in machine learning and natural language processing to transform trading and risk management. However, the deal-making lifecycle remains heavily dependent on human expertise. Despite a decade of progress in automated research, a substantial portion of transaction workflows continues to rely on costly human capital—an inefficiency investment banks are now actively seeking to address.
AI-assisted workflows can now generate first drafts of teasers and memorandums, auto-populate financial profiles from structured data, reduce turnaround time for pitch materials from days to hours, and free senior and junior bankers alike to focus on core issues like building relationships, strategy and insight.
While widespread integration remains nascent, use cases are expanding. Pankaj Harlalka, co-founder of AI-native platform S45, noted that large investment banks are currently “experimenting with AI tools for drafting pitch materials and analysing data”. S45, backed by RTP Global, is an AI-first investment banking platform serving Indian companies across small and medium enterprises and mainboard initial public offerings (IPO).
The shift to AI integration is already yielding pipeline activity, the company claimed. S45 is processing six offer documents and is in advanced discussions with seven to eight companies for mandates, according to another co-founder, Deepank Bhandari.
How will AI change the business model?
The reduction in turnaround time and the potential automation of processes raise questions about the billable value of bankers and the sustainability of advisory fees. Samir Bahl, chief executive of investment banking at Anand Rathi Advisors, argued that the business model will withstand the automation of execution tasks. “Clients don't pay advisory fees for the back-end turnaround time of execution,” he said. They compensate banks for deal origination, marketing and relationship leverage. “Even if absolute fees moderate over time, improved productivity can ensure that profitability remains intact.”
Market participants emphasize that AI is restricted to process-driven tasks. Aman Singh, another co-founder of S45, stated that AI handles document analysis and diligence preparation, while “strategic decisions like mandate selection, deal structuring and transaction judgment always remain with human bankers”.
Addressing concerns that junior bankers might rely on automated models without understanding the maths, Anand Rathi's Bahl noted that “the underlying financial logic still needs to be understood and validated by bankers”, adding that “AI is more likely to augment judgement rather than replace it.”
AI will cause limited job losses while significantly boosting productivity and supporting India's long-term growth trajectory, former HDFC Bank CEO Aditya Puri said at the Mint India Investment Summit in Mumbai on Friday.
To what extent will AI adoption boost revenue?
The same Deloitte report had also predicted that the top 14 global investment banks could boost their front-office productivity by as much as 27%-35% by using generative AI. “This would result in additional revenue of $3.5 million per front-office employee by 2026,” the report had said.
While several large banks that Mint spoke to are already testing their own AI models, none have deployed them on active mandates and thus cannot corroborate the revenue efficiency this has created.
What security concerns does AI adoption raise?
Concerns about data confidentiality and accuracy are still being addressed before such models can be scaled effectively, the head of a domestic investment bank said, on the condition of anonymity.
"Sensitive deal information is typically not shared with public models," Bahl added. Firms rely on secure infrastructure where “confidential inputs are either removed or anonymized”, with expectations that institutions will eventually develop proprietary AI layers trained on internal knowledge bases.
Meanwhile, India's markets regulator is well aware of the use of AI and has established frameworks to manage the resulting compliance risks. The Securities and Exchange Board of India (Sebi) introduced Regulation 16C in its intermediaries regulations in early 2025. Anant Mishra, a partner at law firm J. Sagar Associates, noted the rule dictates that intermediaries remain accountable for AI systems, whether developed internally or via third-party vendors.
“Crucially, the provision eliminates any scope for deflecting liability onto technology providers or opaque algorithms,” he explained. “Irrespective of the technology used, the liability will squarely lie with the intermediary.”
Arun Prabhu, partner at Cyril Amarchand Mangaldas, reinforced that regulated entities must demonstrate “the discharge of obligations of care, including due diligence and fiduciary responsibilities” when using AI tools. This compliance is achieved through judicious data selection and human review to mitigate risk.
“So even if we end up relying completely on AI, the question is who is ultimately ready to sign off on the tool's work and bear the headache of that responsibility,” the investment banker cited above said.