Best AI for Investment Bankers

Investment banking is a profession where the hours are legendary partly because the work is genuinely time-intensive: building models, writing memos, preparing presentations, researching industries, coordinating due diligence. Some of that time is irreducible, the senior judgment on valuation, the relationship management, the negotiation. Some of it is genuinely tedious work that has to be done but doesn't require MD-level thinking.

AI tools can't touch the senior judgment parts. What they can meaningfully accelerate are the writing, research synthesis, and first-draft production parts. That's still real capacity, because the writing and research work can occupy analyst and associate hours that would otherwise be spent in the office until midnight.

This guide covers four tools that fit actual investment banking workflows in 2026. The focus is on the specific tasks where AI produces real output faster and where the quality is high enough to be useful, not just impressive in a demo.

The MNPI problem upfront

Before anything else: investment banking has a confidentiality constraint that overrides most AI tool decisions. Material non-public information can't go into consumer AI tools. If you're working on a live deal, the company name, the transaction structure, the financial projections, and any detail that constitutes MNPI are off-limits for consumer tools like the standard Claude.ai plan or Perplexity.

Most banks have policies about approved AI tools. Check yours. The AI use cases that are safe are ones involving publicly available information, internal research that doesn't involve MNPI, and document preparation that's based on public materials. For live deal work with actual client confidential information, you need enterprise tools with appropriate data agreements.

With that caveat named: here's where AI actually helps.

1. Claude (claude.ai)

Claude is the tool I'd recommend most broadly for investment bankers who want AI help with drafting, research synthesis, and analytical writing.

For deal memos, information memorandums, and management presentations, Claude's drafting quality is high for professional written content. The key is giving it structured input: company background, the strategic narrative you're building, the key financial metrics you want to feature, the comparable transactions you're referencing. With that input, Claude produces a solid first draft of the written sections that a junior banker would have spent several hours on. The draft needs editing, and all factual claims need verification, but the structure and language are good enough to work from.

For comps analysis write-ups, Claude handles the methodology section and the narrative interpretation well. The explanation of why you selected the comparable companies, what the multiple range implies for the target, and how the analysis compares to transaction precedents, these are consistently structured writing tasks that AI does well.

For industry overview sections in pitch books, Claude can synthesize publicly available information about an industry's competitive dynamics, recent deal activity, and key trends. This requires using public information only, but industry background sections are usually built on public sources anyway.

The critical caveats for banking specifically: don't put deal-specific MNPI into the standard Claude.ai plan. And verify every factual claim in AI-generated content before it goes anywhere near a client. The last thing you need is a wrong statistic in an information memorandum.

Best for: Deal memo drafting, IM written sections, management presentation content, industry background, and comps narrative on publicly available information. Pricing: Free tier available; Claude Pro at $20/month; enterprise plan for firm-level deployment with data controls.

2. Gamma

Gamma is the AI presentation tool that builds slide decks from outlines, notes, or brief descriptions. For investment banking, where pitch decks and management presentations are a constant deliverable, Gamma reduces the time from "here's the narrative" to "here's a working deck structure."

The practical workflow is: draft the outline of the pitch in a few bullet points or a short description, run it through Gamma, and get a structured slide deck that you then refine and populate with the actual financial content. Gamma handles the slide layout, the visual hierarchy, and the section structure. You don't have to start from a blank PowerPoint template and figure out where everything goes.

For pitchbook preparation at the beginning of a mandate, when you need to move from initial idea to a client-ready first draft quickly, Gamma is significantly faster than building manually. For full, final deliverables, the output requires substantive editing, real data, and the design polish that major banks expect, but as a starting framework, it's genuinely useful.

Gamma also handles the specific use case of rapid-turnaround internal presentations, where the audience is internal and the content is what matters rather than the design. Strategy update decks, deal team briefings, and committee materials can be produced in Gamma faster than in PowerPoint without the design overhead.

Best for: Pitch deck structures, management presentation frameworks, rapid-turnaround internal presentations. Pricing: Free tier available; Plus plan at $10/month; Pro at $20/month.

3. Perplexity

Perplexity is the external research layer for investment banking work that's based on public information.

The specific use cases are: company background research for prospecting, industry deal volume and transaction trend research, recent public news and announcements about a sector, competitor analysis based on public filings and news, and background on regulatory changes affecting an industry.

All of these are research tasks that analysts do manually, often by searching through multiple sources, reading through articles, and synthesizing findings. Perplexity searches in real time, returns cited summaries, and gets you oriented on a topic faster than the manual approach. The citation structure is important: you need to verify the source before including any fact in client materials, and Perplexity shows you exactly where each claim comes from.

For prospecting specifically, Perplexity helps with building industry overviews, identifying potential acquirers or targets in a sector, and getting current on a company's strategic situation before an initial client conversation. All of this is based on public information, so the MNPI constraint doesn't apply.

At $20/month, Perplexity Pro is worth having for any banker doing significant industry research or prospecting work.

Best for: Company and industry research using public sources, transaction background research, prospecting, and sector overview preparation. Pricing: Free tier available; Perplexity Pro at $20/month.

4. Harvey AI

Harvey AI is the purpose-built professional services AI with enterprise data agreements. For investment banking, the relevant use case is high-volume document review: reading through data room materials, analyzing transaction documents, reviewing representations and warranties in purchase agreements, and extracting key information from large document sets on tight timelines.

Harvey's professional documents training means it understands the structure of acquisition agreements, financing documents, and disclosure materials in a way that general AI tools don't. It knows what to look for in a purchase agreement's material adverse change definition, how to read a working capital mechanism, and where the indemnification caps and baskets are. Feed it a data room, and it can do a first-pass review across hundreds of documents with structured output by document category.

For banks doing significant M&A advisory at volume, Harvey's document review speed on large transactions represents real time savings for associates and vice presidents who would otherwise spend weeks doing the same review manually. The output still requires attorney and banker review, it's not a substitute for professional analysis, but the first pass is substantially faster.

The honest limitation: Harvey is enterprise-priced and not appropriate for individual subscriptions or small advisory shops. The case for it is strongest at bulge bracket and upper middle market banks doing high transaction volume.

Best for: Large investment banks and advisory firms doing significant M&A transaction volume where data room review speed is a competitive factor. Pricing: Enterprise pricing; contact Harvey for current rates.

Where AI fits in the actual deal process

The tools on this list cover different phases of deal work. Most bankers who get consistent value from AI use two or three of them rather than expecting one tool to cover everything.

The combination that works without procurement overhead for most associates and VPs is Claude and Perplexity. That's $40/month combined, requires no IT involvement, and covers research and drafting for most of the written work that doesn't involve live deal MNPI.

For live deal work with actual client confidential information, the tools need enterprise data contracts. That's a firm-level decision, not something individual bankers handle independently.

Frequently asked questions

Can AI help with financial modeling?

For Excel and financial model building, AI code assistants are more relevant than conversational AI. Claude can explain modeling concepts and walk through the logic of a specific calculation, but it doesn't directly edit your Excel model. For model-building assistance, the more useful workflow is describing the structure you're trying to build, having Claude write the formula logic in plain terms, and then implementing it yourself.

Will banks restrict AI tool use?

Many already have. Bulge bracket firms have issued policies on approved AI tools, often restricting use of consumer tools on deal work while evaluating enterprise options. Check your firm's current policy before using any external AI tool professionally. Violations of information security and confidentiality policies in banking are serious; don't assume a tool is approved just because it's publicly available.

Is there a meaningful advantage in using AI early versus waiting for better tools?

For the specific tasks that AI handles well today, research synthesis, deck drafting, and document review, the bankers and analysts who build fluency with these tools now are accumulating workflow advantages that compound over time. Waiting for "perfect" tools means falling behind peers who are already working faster. The tools today are good enough to be worth using; they just require judgment about what to use them for and what to verify.