What Does AI Mean in Investment Banking?
Wondering where those massive AI budgets are going into specifically? In investment banking, AI usually means generative AI (GenAI). That’s software that turns a plain-English prompt into text, a spreadsheet, or a slide. But since client data and deal information are so sensitive, IB firms don’t hand staff publicly available tools like ChatGPT. Instead they've built closed, in-house versions running on their own security systems with only approved internal data.
For instance, Goldman Sachs employees use a tool called the GS AI Assistant, JPMorgan's equivalent is LLM Suite , and Morgan Stanley runs a family of tools including the AI @ Morgan Stanley Assistant for wealth management and AskResearchGPT for its research and banking staff. Outside vendors are also building AI tools for investment banking that plug into Excel, PowerPoint, and internal databases for modeling, data extraction, and document review.
Beyond such GenAI investment banking tools, the next layer IB firms are adopting is agentic AI or agentic processes. Unlike GenAI which simply answers questions, agentic systems act as digital co-workers or agents capable of executing multi-step financial workflows on their own. For example, an analyst can instruct an agentic system to locate a specific set of target cross-border acquisitions, pull their financial histories, convert currency fluctuations, and flag compliance risks. That is where IB technology is heading.
How Analysts Work Today
Though AI is changing and will change a lot in IB, especially for junior staff, the traditional investment banking analyst workflow is a manual, labor-intensive one with tight deadlines and senior banker requests. Analysts spend the bulk of their 80-to-100-hour workweeks juggling between these main tasks:
- Financial Modeling (Excel): Manually sifting through public filings (10-Ks) and databases or terminals like Bloomberg to pull the data behind comparable company analysis (comps) tables or building other valuation models.
- Confidential Information Memoranda (CIM) and Pitchbook Creation (PowerPoint): Formatting 30-to-40-page marketing presentations, adjusting font sizes, aligning graphics, and updating industry charts.
- Administrative Due Diligence: Sorting through virtual data rooms, scouring company registries, checking contracts and financials line by line, and summarizing endless regulatory filings.
In most cases, senior feedback arrives late in the afternoon forcing analysts to routinely spend their nights manually inputting revisions, updating financial models, and cross-checking formulas to ensure error-free deliverables by morning. As such, the primary value of a junior banker has always been measured by their speed at research and data entry as well as their attention to detail during repetitive formatting tasks.
Tasks AI Can Already Support
If you look keenly at those tasks IB analysts perform, you’ll notice they’re quite repetitive, and that’s exactly the kind of work generative AI is good at. So, most of the tasks AI can already support today is much of what analysts do including pitch book development, financial modeling, market research, and due diligence.
Below is an overview of how AI in investment banking is automating these repetitive tasks and reducing manual busywork.
Pitch Book Development
Creating pitch books is traditionally one of an analyst's most labor-intensive tasks. But AI can now help with that. Automated systems can generate deal summaries, draft rationale slides that clearly outline why an acquisition makes strategic sense, and build presentation-ready slides by applying strict corporate templates and color palettes.
In one internal demonstration, JPMorgan's LLM Suite reportedly produced a full first draft of an investment banking presentation, complete with presentation-ready slides, recent news, earnings, and a peer comparison, in about 30 seconds. That's a clear example of accelerating deal execution and cutting real turnaround time.
Financial Modeling Automation
Another area where AI is reducing manual busywork is financial modeling, and most of the AI tools for investment banking being built are in this category. Generally, financial modeling automation tools assist in updating financial models with fresh data automatically instead of manually typing historical numbers into a spreadsheet.
For instance, a purpose-built tool called Rogo plugs directly into Excel and PowerPoint to help analysts roll forward complex models and catch formula errors. It has reportedly become standard at some firms including Lazard, Jefferies, and Moelis. Another tool, Daloopa, focuses specifically on extracting structured data from unstructured documents like company filings, cutting out hours of manual entry and the transcription mistakes that come with it.
Market Research and M&A Deal Sourcing
Research and writing tasks are another example of what AI can already do well in IB. Morgan Stanley's AskResearchGPT lets banking and research staff search tens of thousands of the firm's research reports in plain language instead of digging through them by hand. A separate tool, AI @ Morgan Stanley Debrief , automatically turns client meeting notes into summaries and follow-up emails.
Goldman's GS AI Assistant is used firmwide to summarize long documents, draft first-pass content, and even translate research into other languages for clients abroad. Besides, AI is good at monitoring global markets, news feeds, and alternative datasets to identify potential buyers or acquisition targets. This means it can generate M&A pipelines in hours instead of weeks.
Due Diligence & CIMs
During initial deal screening, advanced document intelligence tools are good at extracting structured data from unstructured documents. That means the tech instantly turns messy bundles of client contracts and invoices into organized, sortable data tables.
As a deal moves forward, junior teams can use specialized AI tools, like Hebbia , for CIM triage (Confidential Information Memorandum). Instead of reading cover-to-cover, analysts can instantly isolate key risk clauses or revenue breakdowns across the document.
Then for deep-dive reviews, AI ingests entire Virtual Data Rooms (VDRs) to locate specific metrics and generate automatic citations. This eliminates the need for analysts to manually scan thousands of pages, allowing them to synthesize their diligence findings in a fraction of the traditional time.
What AI Still Cannot Do Well
Though AI handles such repetitive, data-heavy grunt work well, and there's already progress in M&A workflow automation, it still falls short in some areas. That’s mostly due to IB's relational nature, strict regulatory demands, and the technology's own limitations like errors and hallucinations.
Here’s an overview of what AI still cannot do well in investment banking:
- Delivering Client-Ready Materials Alone: Accuracy is a big deal in IB yet AI tools are prone to model bias and AI hallucinations. So, it requires strict human oversight or human judgment to evaluate complex data, facts, or AI output. Analysts must check source citations or audit trails before trusting any number.
- Conducting Negotiations & Handling Client Relationship Management: Investment banking is a relationship business built on trust. AI cannot read the room during a tense boardroom negotiation, build a long-term rapport with a founder looking to sell their life’s work, or gauge the emotional hesitation of a CEO during a hostile takeover.
- Performing Evaluation & Risk Assessment Independently: AI can flag risks that show up in the data like leverage, litigation history, or customer concentration. But it can't weigh the unquantifiable ones, like reputational fallout or how a market will react, the way a banker's pattern-recognition from past deals can.
- Guaranteeing Data Protection and Compliance: Since data privacy is non-negotiable, banks strictly prohibit public models to prevent insider trading leaks. Instead, firms require isolated, enterprise-grade protection. Every automated workflow must also be fully auditable to withstand intense SEC and FINRA scrutiny.
How the Analyst Role Is Changing
If AI or investment banking automation is taking up the production side of the work, the natural next step is for the analyst's role to shift toward human oversight. More like reviewing and directing that production instead of doing it from scratch. Instead of spending time building a comps table from zero, more time will be spent checking whether the comps table AI built actually makes sense.
That directly changes the metrics and value of junior bankers. As an analyst, your primary value to a deal team will no longer be your speed at data entry or your ability to perfectly format a slide. It will instead be what you do with the information once a system aggregates it for you like identifying anomalies in the numbers, stress-testing valuation assumptions, and formulating strategic advice for the client. This means IB analysts are expected to focus on high-value strategic priorities much earlier in their careers.
Skills Future Investment Banking Analysts Need
As the IB analyst role changes due to investment banking automation, the specific things that make someone good at the job are also shifting. Here is how to stay relevant by mastering essential AI investment banking skills
- Critical Thinking & Skepticism: Analysts must transition from creators to expert editors. Since AI models can present subtly incorrect data with absolute confidence, applying human judgment to AI outputs is non-negotiable. This includes identifying what is missing or logically inconsistent.
- Thorough Research & Fact-Checking: Verifying facts against foundational sources must become a core habit rather than an afterthought. Analysts must aggressively track down source citations and audit trails to catch hallucinations and false references before they reach clients.
- Deep Business Model, Deals, & Industry Fluency: AI can format slides, but it lacks sector intuition. Human expertise is required to evaluate supply chains, map competitive moats, and ultimately determine which businesses truly belong in a comparable company analysis (comps) set.
- Technical Fluency: You don't need to become a software engineer, but you do need to structure complex queries, guide agentic systems, and prompt well enough that the output is usable on the first pass, not a rough draft you end up rebuilding.
Will AI Replace IB Analysts?
No, AI will not replace IB analysts, at least not in the foreseeable future. In the short to medium term, banks are not eliminating the analyst position entirely. They still need a pipeline of junior talent to apply human judgement to AI outputs and be responsible for the work. Also they need them to learn the business from the ground up so they can eventually step into leadership roles as Vice Presidents and Managing Directors. If you remove the entry-level tier, you destroy the future leadership of the firm.
However, AI’s impact on investment banking jobs may revolve more around headcount reductions. One analyst equipped with enterprise AI tools can efficiently handle the output that used to require two or three analysts, meaning the hiring market may become more competitive. So in the near and medium term, what you can expect is a redefinition of what the job requires and a shrinking of how many people it takes to do it. This might soon mean smaller analyst classes at some firms or similar classes but more deals.
What Does This Means for Candidates?
If you are currently applying for internships or full-time analyst roles, you may start wondering whether investment banking is still a good career to pursue. The short answer is yes. But getting an offer may be more competitive than it used to be, simply because smaller entry-level classes mean more people competing for fewer seats at some firms. Standing out will require demonstrating real AI investment banking skills.
At the same time, the fundamentals haven't gone anywhere. Understanding accounting, valuation, and how a model actually works still matters, arguably more than before, because the job now often involves checking someone else's, or something else's, output rather than building everything from a blank page. Reviewing work well is harder to learn than producing it, since you need to already know what right looks like before you can catch what's wrong.
Also, the future of the investment banking analyst role may not involve lesser hours as a result of automating the grunt work. The time saved on one task may get absorbed by more deals or higher expectations rather than turning into space for anything else.
Conclusion
AI is not killing the investment banking analyst role, but it’s changing the day-to-day tasks from producing materials to supervising AI that produces them. As a result, the future of the investment banking analyst belongs to the ones treating AI as a tool worth mastering, not a threat to ignore or a shortcut to lean on blindly. This means adaptability is key. After all, IB firms are already investing heavily in this tech.
You can expect a short-term disruption or transitional period with potential job reductions and increased competition, but also opportunities for those who can leverage AI effectively. Most importantly, AI won't replace everything in a banker’s work. Operational inefficiencies, relational requirements, and the need for human judgment in complex deals and AI outputs will ensure that some roles remain intact.
