AI outmoding Investment/Corporate Banking?
Lately, I've been contemplating the impact of recent advancements like Bard and ChatGPT, particularly in relation to the banking industry. The prevailing argument I often hear is that banking relies heavily on personal relationships and face-to-face interactions, and while that may hold true for now, it's worth examining the nature of these interactions. Speaking from my experience in corporate banking, let's delve into it further.
During a lunch meeting, my managing director (MD) engages with a corporate treasurer or CFO (the client) who expresses a need for financing an acquisition. Armed with that information, my MD collaborates with our team and product partners to determine the most suitable structure for the facility, be it a term loan, revolving credit facility (RCF), or a straightforward bond issuance. However, envision a future where the client could employ a "Treasurer AI" to analyze their company's real-time financial health, cash position, and capital structure, effectively determining the optimal form of financing. Similarly, the bank could utilize an "AI MD" to instantaneously structure the debt in the most capital-efficient manner. Suddenly, an industry traditionally driven by personal relationships becomes an interaction between two artificial intelligences. This scenario isn't far-fetched, considering that AI has already played a role in lending decisions for the past decade. The entire credit process can be automated.
The example I provided is just one illustration. An AI MD could conceivably generate a comparable analysis and perform precedent runs within seconds, presenting the information as binary code to an AI CFO, who could then chart the best path forward for the company.
These are simply musings on a Friday night from my perspective, and would love to hear your thoughts. Maybe arguments on what parts of banking would be more “AI-proof”?
Also, as a disclaimer, I’m not the best writer so I fed my original draft of this post to AI and had it correct any grammatical mistakes. This final product you read is much better than what I wrote
.
Great. Training AI in 15 years. Exactly why I got into this industry!
.
Strongly disagree with this if your reasoning is that the technology isn't there yet - there are already AI platforms that are used for personal financing which evaluates everything from your portfolio, investment objectives, market conditions, etc to recommend tailored investment advice.
That being said, a large bottleneck I can see with creating a model that could actually output feasible M&A advice is the amount of data available on this used to train a model. You would not only need an exorbitant amount of data, but high quality, comprehensive information about past deals to train a model that is capable and reliable enough to provide sound advice. Considering the relatively recent timeline of the PE / M&A boom, this would be logistically very difficult to gather.
But back to the main discussion, I could not see AI replacing the skills of a higher-level investment banker where relationships are core to their value prop. Not only because relationships are exclusive to humans (at least for now lol), but also because companies need/want accountability. If things go south with a deal and the advice was purely provided by an AI, who do you have to blame, and whose reputation is on the line?
However, I can absolutely see a lot of the junior bankers' responsibilities being automated in the next 3-4 years. Analysts are not doing any critical thinking which leads to decision-making, nor do they have relationships that drive revenue.
So far, tech has proven to be great at automating mundane, repeatable tasks. If you think about an Excel spreadsheet, it is computing millions of different things at once. I assume this used to be some analysts' job to do the math with calculators and write it down on paper for their associate to create analysis with. I think it's the same way with AI but to a greater extent. It may automate the mundane analyst responsibilities of buyer tracking, finding comparables, number crunching, etc. But when it's all said and done, you need a human to pull the trigger on making large decisions.
Good job, Bard.
What makes you qualified to make this statement? Did you use ChatGPT and read news about GPT-4 and then decided this with actual industry experience? Lol
Quibusdam quia aliquam labore perferendis. Eaque quidem quia sint commodi et sed voluptatibus. Quasi necessitatibus voluptate ipsum iste autem vitae natus suscipit. Consequatur minus esse ut ut sapiente cupiditate necessitatibus.
Sit soluta sunt laudantium neque cumque reiciendis impedit. Sit cumque et voluptatem non qui. Dolores eum rerum dolore aperiam minus. Quibusdam quo eum voluptatem voluptatem saepe occaecati.
Aut distinctio delectus aut labore. Est enim necessitatibus est ducimus illum amet. Blanditiis corrupti autem sed architecto in est quia dolores. Natus iusto praesentium explicabo cumque ea. Perferendis sunt qui omnis sit qui qui dolore.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...