The AI Hype is Real. But So is the BS.
Will artificial intelligence kill your Wall Street career before it even starts? It’s the question echoing through business schools and analyst bullpens alike.
The hype is everywhere, from headlines declaring that “the worst part of a Wall Street career may be coming to an end” to CEOs boasting about AI drafting nearly entire IPO prospectuses in minutes.
But let’s cut through the noise.
In this article, we’ll break down what’s really happening with AI in investment banking.
Let’s start by examining whether investment banking is truly becoming automated or if this is just another case of Silicon Valley salesmanship gone wild.
Is Investment Banking Becoming Automated?
There’s no doubt that automation is creeping into investment banking, an industry historically known for armies of analysts manually churning through data and PowerPoint slides.
On the one hand, we see clear signs of progress. Banks are experimenting with generative AI to speed up research, due diligence, and even document drafting.
A recent Bain & Company report highlights that use of gen AI in M&A processes grew from 16% of companies in 2023 to 21% in 2024, and is expected to surpass 50% by 2027.
In other words, adoption is climbing. Nearly 80% of those using AI in deal-making say it cuts manual work, and over half claim it accelerates deal timelines.
So yes, parts of investment banking are becoming automated.
But note the phrasing: parts, not the whole. Automation in IB is currently surgical, not wholesale.
Banks are using AI to augment their teams, not to replace entire deal teams with robo-bankers.
Much of the change so far has been focused on making the existing work faster and less painful, rather than fundamentally changing what the work is.
In the next section, we’ll get concrete with examples of how AI is actually used in IB today, beyond the buzzwords.
AI Use Cases in Investment Banking
It’s time to go beyond theory and see how AI is used in practice by investment bankers today.
Research and Market Intelligence
Investment banking runs on information.
Traditionally, a junior banker might spend days scouring SEC filings, equity research reports, news, and internal databases to understand a client’s industry. Now, AI-powered research tools are supercharging this process.
For example, Morgan Stanley recently launched AskResearchGPT, an internal generative AI assistant that lets its bankers instantly search and summarize insights from over 70,000 proprietary research reports.
Instead of manually flipping through PDF reports, bankers can ask a question and get a distilled answer with sources, in seconds.
Even S&P Capital IQ – a staple database in finance – rolled out a generative AI feature called Document Intelligence.
It can summarize lengthy 10-Ks or earnings call transcripts and answer questions about them via a ChatGPT-style interface.
What used to require an analyst pulling an all-nighter reading can now be done by an AI in a coffee break’s time.
Slide Creation and Document Drafting
Ask any junior banker their least favorite task, and many will say “turning comments”; i.e. the endless edits and formatting fixes on pitch decks and memos.
Here’s where AI is chipping in. Generative AI is now drafting the first versions of banker documents.
According to a report by Mckinsey, one leading investment bank built a genAI tool to help analysts write first drafts of pitch books, generating many of the standard slides automatically, a move that cut pitch book prep time by ~30%.
Similarly, Microsoft’s new Copilot (an AI assistant in Office 365) can draft entire PowerPoint presentations or Word documents from a simple prompt.
Bankers are already testing Copilot to do things like create an investment memo template, rewrite tedious prose, or suggest charts for a deck, all in seconds.
In fact, Goldman Sachs is reportedly developing an AI tool that can transform a long-form PowerPoint into a polished S-1 registration statement (the dense legal document for IPOs) in less than a second.
Financial Analysis and Modeling
This is emerging, but worth noting. AI isn’t yet clicking the buttons in your Excel models, however, it’s starting to assist around the edges.
For instance, JPMorgan has a proprietary AI dubbed IndexGPT in a patent filing, aimed at helping generate stock portfolios or analysis for clients.
More immediately, Copilot in Excel can write complex formulas, build pivot tables, or run what-if scenarios based on a simple request.
Imagine telling Excel Copilot, “Hey, pull the last 5 years of EBITDA for our comp set from CapIQ and chart the median – and highlight any outliers.”
We’re not quite there in fully automated form, but we’re close.
Can AI Fully Replace Investment Bankers?
Despite impressive advances, the idea that AI will fully replace investment bankers is, at best, wildly premature, and at worst, plain fantasy.
Investment banking isn’t just spreadsheets and pitch decks; it’s an inherently human business in many respects.
Here’s why the human banker isn’t an endangered species (yet):
Client Relationships and Trust
Investment banking is often described as a “relationship business.” CEOs and CFOs want a trusted advisor, someone who understands their business and whom they can hold accountable.
They are not going to entrust multi-billion-dollar strategic decisions to an AI, no matter how advanced, without a human in the loop.
Can you imagine a board meeting where the lead advisor is a laptop running ChatGPT?
Neither can I.
Clients hire bankers for judgment, discretion, and the sense of security that comes from having a human accountable if things go sideways.
Trust is earned by humans, not algorithms.
Complex, Unstructured Problems
A lot of what junior bankers do is structured and can be automated – spreading financials, summarizing slides, etc.
But the higher-level work in IB is messy and unstructured.
Figuring out how to salvage a deal when two CEOs hate each other, or devising a completely new financing structure during a market crisis, or simply having a creative idea that a client hasn’t considered – those tasks don’t have clear rules or huge datasets to train on.
AI is great at pattern recognition from past data; it’s less great at true creativity or handling one-off novel situations.
Regulatory and Ethical Constraints
Banks also have to be extremely careful about deploying AI.
There are compliance and legal considerations around client data, which is why firms like JPMorgan barred employees from using public ChatGPT and built their own firewall-protected versions.
Moreover, AI models can “hallucinate” (i.e., make stuff up) or use biased data, which in a regulated setting could be disastrous.
Imagine an AI that inadvertently uses confidential information in a pitch, or suggests a course of action that runs afoul of securities law.
Ultimately, a human has to be accountable to regulators and clients, and the bank will not sign off on AI making decisions without human review. The technology is being used as an assistant, not the decision-maker.
What Roles Are Most at Risk?
While AI won’t send all bankers packing, it will reshape the role of certain bankers – especially at the junior levels.
If you’re a student or new analyst, you’re probably wondering, “Am I automating myself out of a job by learning this stuff?”
Here’s the nuanced view on which roles and tasks are on the chopping block and which are relatively safe:
Analysts and Associates (Junior Bankers)
Sorry, team – a lot of your traditional work is exactly what AI is learning to do.
First- and second-year analysts historically spent their days (and nights) building financial models, putting together decks, doing industry research, and proofreading documents.
These task-level activities are prime targets for AI.
Instead of ten analysts combing through data for a market update slide, maybe one analyst with an AI tool can do it in a morning.
Your training as a junior may involve overseeing AI outputs and focusing on higher-level analysis rather than manually doing all the drudge work.
That’s exciting – but also daunting, because it compresses the learning curve.
Middle-Office and Support Roles
Think of roles like research associates, data room coordinators, or even some IT support functions in deals.
If your job is to pull data, generate reports, or prepare standard documents, AI can probably handle a lot of that.
We’re already seeing research departments use AI to produce first drafts of reports, with humans editing and adding insight.
In due diligence support, one person with an AI tool can potentially do the work that used to require a small army sorting documents.
If your job can be described as “take info from point A, reformat it, and send to point B,” an AI can learn to do that.
The key for these professionals will be to upskill, learn to work alongside AI and focus on interpretation, quality control, or specialized knowledge that the AI isn’t trained on.
Senior Bankers (VPs, Directors, Managing Directors)
In general, senior client-facing bankers are least directly threatened by AI, but their roles will evolve too.
They’re not at risk of being replaced by a bot, but they will be expected to leverage AI tools to be more productive.
A Managing Director in the future might cover more clients or execute more deals simultaneously, because with AI, a smaller team can accomplish the same amount of work.
Specialists (and New Tech Roles)
Interestingly, we may see new roles emerge that didn’t exist before.
Banks are already hiring AI specialists and prompt engineers, people who know how to fine-tune models or craft prompts to get the best results from AI.
If you have a dual passion for finance and tech, there could be a niche for you in developing and overseeing the bank’s AI tools. Also, certain specialists like quants and technologists will collaborate more with bankers, blurring lines.
For example, a deal strategist who knows AI could work alongside coverage bankers to identify acquisition targets using AI-driven analytics that a pure banker might miss.
Expertise in both finance and AI will be a highly valuable combo.
Which Investment Banks Are Actually Using AI?
By now you might be wondering, “This all sounds great in theory, but are big banks really using these AI tools, or just talking about them?”
The reality is some firms are much further ahead than others. Here’s a quick tour of how a few notable investment banks are embracing AI:
Morgan Stanley
Among Wall Street banks, Morgan Stanley has been quite aggressive in AI adoption.
As mentioned, they launched AskResearchGPT in late 2024 for their investment banking and markets teams.
This internal GPT-4-powered chatbot lets bankers query the firm’s massive research library and get instant answers.
They reportedly have dozens of AI projects in the pipeline, ranging from automating parts of trading to improving how they match potential buyers and sellers in M&A situations.
Goldman Sachs
Goldman has taken a very high-profile approach to AI.
CEO David Solomon is openly bullish, and the firm’s top partners have spoken about AI as the next major driver of efficiency Goldman has assembled an internal “arsenal” of AI tools – some off-the-shelf, some built in-house.
For example, Goldman’s AI sidekick (as some employees call it) is used by certain teams to help with tasks like coding (yes, even bankers need to automate Excel sometimes), generating analytics, and drafting content.
JPMorgan Chase
JPMorgan has been a bit quieter publicly but is doing a lot behind the scenes. They famously banned employees from using ChatGPT early on, citing confidentiality, and then rolled out their own in-house LLM platform for employees called LLM Suite.
This is essentially a JPMorgan-trained ChatGPT-like tool that 50,000 of their staff can use for general productivity – writing emails, summarizing documents, generating code, etc.
What’s the Future of AI in Investment Banking?
Projecting the future is always tricky.
But we can make some educated predictions about how AI will shape investment banking in the coming years:
Faster Deal Cycles
With AI handling preparation work, the overall timeline of deals could compress. We’re already seeing early evidence of faster due diligence and faster document turnarounds.
In the future, a sell-side M&A process that used to take 6 months might close in 4, because the info is organized and analyzed more rapidly and bidders can move quicker.
IPOs could be filed faster. This could mean more deals get done in a given year, or at least bankers can handle more concurrently.
Leaner Teams, Different Skills
The classic pyramid model might evolve.
We may see somewhat smaller deal teams where each person is empowered by AI to do more. For example, instead of 3 analysts on a big deal, maybe 1 analyst with good AI tools can do the work. Those tools might be considered part of the “team.”
At the same time, the skills banks value will shift. Coding or at least scripting ability might become a standard expectation.
The ideal banker of the future might be a bit of a hybrid: financially savvy, great with people, and capable of wielding AI tools effectively.
Training programs may start to include modules on how to prompt ChatGPT or how to verify an AI-generated analysis.
New Tools Integrated Seamlessly
The AI tools we described will likely not remain standalone for long.
They’ll get integrated into the software bankers use daily. We can expect our trusty Excel, PowerPoint, Word to have ever smarter “autopilot” features.
Bloomberg and Refinitiv might integrate chat-based interfaces to retrieve data or do analysis from their terminals.
So the banker of the future might chat with their computer to get work done: “Find the top 5 comps for Company X and make a slide comparing their EBITDA margins” – and within minutes have something close to ready.
This doesn’t eliminate the banker; it changes their workflow. You become more of a director or editor than an assembler of content.
Closing: Should You Be Worried? Or Excited?
If you’re an aspiring analyst or a junior banker, how should you feel about AI’s growing role in investment banking?
My honest take: a bit of both worry and excitement is justified – but lean towards excitement. Here’s why:
Why You Might Worry
Change is always uncomfortable.
AI in banking means the job you thought you were signing up for will evolve.
Some traditional paths might become less available.
Also the lower-skill floor is rising – meaning, you’ll need to bring more to the table to succeed. That can be intimidating.
Why You Should Be Excited
Frankly, a lot of the work AI is taking over sucks to do.
It’s the tedious stuff we all complained about.
If you actually get to spend more time thinking about big-picture problems, interacting with clients, or honing persuasive skills, and less time resizing text boxes or copy-pasting numbers – that makes the job more fulfilling.
Also, being young and adaptable is an advantage. You can become the go-to person for new tech in your team.
Senior folks might resist or be slow to learn these tools, so juniors who master them can shine
So, don’t be afraid of AI; get good at it.
Learn to use AlphaSense to do rapid research. Play around with ChatGPT or Copilot to see how it can draft slides or memos.
If your firm has an internal AI tool, volunteer to pilot it.
By mastering these, you make yourself more valuable, not less.
You’ll free up time to ask the interesting questions and make insightful recommendations, which is what good bankers are supposed to do.