Does Corporate and Investment Banking Need Improved Efficiency from AI?

In the second of our Generative AI series, expert Kayode Odeleye discusses what the future could have in store for us with this new technology.

What happens on Wall Street stays on Wall Street. Like their gambling counterparts, investment banking operates under an unspoken rule of omerta. This code of silence was shattered in March 2021 by a group of junior Goldman Sachs analysts. A leaked survey presentation from 13 TMT analysts revealed their complaints and suggestions to senior managers. The analysts reported facing inhumane conditions, including 100-hour workweeks, abuse from colleagues, and severely impacted mental health. They requested the enforcement of one work-free day per week and a cap on 80-hour workweeks.

Junior staff in finance have long accepted the trade-off: work extremely long hours and in return, receive excellent compensation. Unfortunately, these demanding hours are often spent on tasks that do not add direct value to clients. Analysts invest significant time in researching, creating presentations, and building financial models. The industry is also notoriously inefficient with lots of manual data entry and a resistance to change.

Generative AI Blog Post Series 2.png

What Can Technology Do to Improve Processes?

With the recent advancements in generative AI, it is inevitable that senior finance professionals must be asking at least one of these two questions:

  1. Can we reduce our cost base by replacing junior bankers with technology?

  2. Can we give them better tools so they achieve more while having a better quality of life?

The answer to these questions will depend on how well Generative AI tools can perform the jobs being done by junior bankers today. More importantly, it will also depend on how quickly these capabilities improve.

So how much of what corporate and investment bankers do today can AI help with? Let's run ChatGPT through its paces to find out.

Input: Prompt the system to read screenshots of a public company's financial statements to generate a financial model and identify issues with the capital structure.

blog pic1.png

Output 1: Create an Excel file from screenshot

blogpic2.png

ChatGPT took just 30 seconds to read the image and convert the numbers into a spreadsheet. As seen below, the Excel file had two separate tabs, one each for the P&L and Balance Sheet. I did a comparison with the original file and the conversion was 100% accurate!

1 (1).png

Verdict: the accurate, well laid out but unformatted file is the level expected of an investment banking intern with limited training. However, it would take the most experienced numbers cruncher at least 10x longer to complete the same task. AI wins, hands down.

Output 2: Analysis of the financial statements and recommendations

2.png

Verdict 4/10: the observations are obvious from the balance sheet but requires some level of training. A random university educated junior analyst may be able to come up with same without any hand holding. However, the initial suggestions require some knowledge of capital markets and are quite thoughtful. When provided with additional information "A large part of their debt is in form of bonds, does that change your recommendation?", "it" comes up with this:

blogpic3.png

AI is here to stay. As these experiments as demonstrated, existing, generally available generative AI tools are at least as efficient in some Banking / Finance tasks as a lightly trained financial analyst. The consideration is that these AI tools are getting exponentially better [Moore’s Law for AI]

The conundrum facing leaders is two-fold: make quick investment decisions on AI tools or wait to see whether it’s all hype and will soon blow over. There are risks to both extremes. On one hand, investing in new AI tools could be premature (remember blockchain a few years ago?) and all the investments go down the drain if this ends up being another flash in the pan. On the other hand, financial institutions risk being left behind if competitors gain significant productivity and other advantages from getting onboard the AI train early.

There are no easy answers on navigating this delicate balance except dedicating time and energy to thoroughly understanding capabilities as it relates to each organisation.

In less than two-years, game-changing generative AI tools have gone from zero to wide adoption faster than any other consumer tool in history. Organisations, including banks and investment firms, are grappling with decisions on how to integrate these tools to enhance productivity while protecting their organisations from inherent risks.

On the 17th of September, Euromoney Learning will be running a webinar called Innovate or Stagnate: Top 5 Generative AI Use Cases in Corporate Finance with the author.

Kayode Odeleye is an investment banker and tech startup founder will help executives make sense of the fast-changing landscape and identify major risks and opportunities involved with implementation of generative AI in their institutions; and will share the top 5 generative AI use cases in corporate finance.

You can watch Kayode in action in our popular on demand webinar here.

If you haven’t already, don’t forget to read our previous blog post in this series, How AI is Transforming Corporate and Investment Banking

Gift this article