HIGH RISK ■ Finance

Will AI Replace Investment Banker?

The analyst factory floor, yes; the managing director, no. Banking's pyramid is built on armies of juniors doing exactly the work that generative models produce in seconds.

58%

AI models deals and runs valuations. But closing requires charm, steak dinners, and a firm handshake.

Our AI replacement risk score — how we score jobs

Why Investment Banker scores 58%

The industry's structure explains its exposure. Analysts and associates spend brutal hours building three-statement models, running comparable company and precedent transaction analyses, formatting pitchbooks, updating CIMs, populating data rooms, and reformatting the same slide because a VP wanted a different shade of blue. Vice presidents manage that output; directors and managing directors originate business — calling CEOs, cultivating boards, pitching for mandates, and negotiating the terms that close a deal.

Everything below the VP line is squarely in automation's path. Models generate DCFs and comparables from filings, assemble pitchbooks in house format, draft memoranda, summarize diligence documents, and perform first-pass document review far faster than a 22-year-old at 3am. Banks have publicly moved to smaller analyst classes and have said openly that AI changes junior hiring plans. This is not speculative displacement; it's a visible change in intake. Our 58 reflects a profession where the majority of headcount sits in the automatable tier.

Origination is a different animal. Mandates come from relationships built over years — knowing which family-owned business is finally ready to sell, having the CEO's trust when they're deciding whether to defend against an activist, being the person a board calls at midnight. Negotiation dynamics, reading the other side, managing a seller's emotions about a company they founded, and structuring a deal that both parties can accept are human trades. Regulatory sign-off, fairness opinions, and fiduciary duty also demand accountable humans. The likely equilibrium is a much flatter pyramid: far fewer juniors, similar numbers of senior bankers — which raises an awkward question about where the next generation of senior bankers gets trained.

Which Investment Banker tasks can AI automate?

Building valuation models, comparables, and precedent transaction analysesHIGH
Producing and formatting pitchbooks and marketing materialsHIGH
Document review and data room preparation during diligenceHIGH
Managing a live deal process across lawyers, buyers, and managementMEDIUM
Originating mandates through long-term client relationshipsLOW
Negotiating terms and reading the counterparty in the roomLOW

Automatability: our editorial assessment of current and near-term AI capability

When will it happen?

Already underway. Major banks have signalled smaller junior classes and deployed internal AI tooling for modelling and document work, and analyst headcount is the visible casualty. Expect the pyramid to flatten sharply through 2030. Senior origination roles remain secure well beyond that, but the apprenticeship model that produced them is breaking — a structural problem the industry has not solved.

How to stay ahead

  • 01Get client exposure absurdly early; the analyst who is only an output machine has the shortest runway.
  • 02Specialize in a sector deeply enough to have a view, not just a model.
  • 03Master AI tooling to produce more with fewer hours, and make that visible to the people staffing deals.
  • 04Consider adjacent paths — private credit, corporate development, PE operating roles — where relationships develop faster.

Investment Banker & AI: common questions

Are investment banking analyst jobs going away?

They're shrinking meaningfully. The analyst role exists to produce models, decks, and diligence materials, and AI does all three quickly. Banks have already begun trimming junior classes and pointing to automation as a reason. The role won't vanish entirely, but expect fewer seats, higher expectations, and less tolerance for people whose value is output volume.

Can AI actually build a valuation model?

Yes, competently, from filings and market data — including comparables and precedent transactions. Where it needs supervision is in judgment calls: which comparables genuinely reflect the business, what adjustments the management projections deserve, and how to defend the assumptions when a counterparty attacks them. Model construction is commoditizing; model defensibility isn't.

What part of banking is safest?

Origination. Winning mandates depends on long-cultivated relationships with CEOs and boards, sector credibility, and being trusted at the moment a company decides to sell or defend itself. Negotiation and deal management also resist automation because they involve reading people under pressure. The safest bankers are the ones who bring in business, not the ones who process it.

Is banking still worth entering as a career?

It can be, but the calculus has changed. The traditional deal — endure two brutal years of modelling to earn access to the senior track — is weaker now that the modelling itself is being automated and junior classes are smaller. If you go in, optimize aggressively for client contact and sector expertise rather than technical output, and treat exit options as a live plan.

Related jobs