AI agents are now negotiating M&A term sheets while bankers sleep, and the banks that figure this out first will eat everyone else's lunch.
The Summary
- Key Institutional Bank, McKinsey, and a16z executives discussed how AI is reshaping dealmaking infrastructure on Bloomberg Deals
- Banks are deploying AI agents to handle due diligence, contract analysis, and deal structuring work that used to take armies of analysts weeks to complete
- The shift isn't just efficiency gains, it's a fundamental reorganization of who does what in high-stakes finance
The Signal
The institutional banking world is quietly automating its most lucrative work. When a banking president, a McKinsey senior partner, and an a16z general partner show up together to talk AI dealmaking, they're not discussing theory. They're describing what's already happening behind NDAs.
Randy Paine from Key Institutional Bank joined Amit Garg from McKinsey and Angela Strange from Andreessen Horowitz to map the current state. The conversation centered on how AI agents are taking over the grunt work of M&A: combing through data rooms, flagging regulatory issues, running financial models, even drafting initial term sheets. This isn't speculative. These are the people building and deploying these systems right now.
The a16z perspective matters here because they've been funding the picks-and-shovels companies building agent infrastructure for finance. McKinsey is in every boardroom telling executives to automate or die. And Key's institutional bank is actually doing the deals. When those three perspectives align, it means the future already arrived, it's just not evenly distributed yet.
What makes this different from previous waves of financial automation is agency. These aren't just faster calculators. They're systems that can read a hundred contracts, identify unusual clauses, cross-reference them against regulatory databases, and surface decision-ready insights without human handholding. The junior analyst job isn't getting easier. It's getting eliminated.
The Implication
If you work in finance, ask what parts of your job involve pattern recognition, document review, or data synthesis. Those parts are getting automated fast. The work that survives will be relationship management, strategic judgment, and navigating the messy human parts of deals that agents can't touch yet. Banks that build agent teams faster than competitors will close more deals with fewer people. Watch which institutions start shrinking analyst classes while deal volume stays flat or grows.
Source: Bloomberg Tech