The useful question about AI in investment banking is simple: where can it improve the quality and speed of the work?
Banking combines very different kinds of work, ranging from repetitive information processing to judgment under extraordinary pressure.
Automation suits the repetitive work. Senior professionals remain responsible for judgment and outcomes.
Replace the work that exists because information is fragmented
Bankers spend enormous amounts of time gathering information that already exists somewhere: company details, ownership, transactions, comparables, buyer activity, lender preferences, sector developments, and prior conversations.
The information matters, and technology can handle much of the repetitive collection.
Technology should be able to monitor a defined universe of companies, detect meaningful changes, maintain comparable-company sets, update counterparty intelligence, and prepare a first-pass briefing before a senior person begins the analysis.
It should also reduce the administrative drag around meeting notes, pipeline updates, document organization, and process coordination.
Keep accountability with the banker
A model can produce an answer, while the banker carries responsibility for the advice.
When an owner is deciding whether to sell the company they spent twenty years building, the advice has consequences beyond a spreadsheet. Someone needs to understand the assumptions, challenge the analysis, explain the trade-offs, and stand behind the recommendation.
The most valuable judgment appears precisely where the data stops being complete.
That includes understanding whether a buyer is serious, whether management will survive a transition, whether a financing structure leaves enough room for a downturn, and whether the highest bid is actually the best outcome for the owner.
Trust moves the transaction
Transactions move because people believe one another.
A founder has to trust that their adviser understands what is at stake. A buyer has to trust the information. A lender has to believe the downside case. Management has to believe the process is being handled carefully. When something goes wrong, the parties need a person who can absorb the tension and keep the transaction alive.
AI can help someone arrive better prepared. Accumulated trust allows the difficult conversation to happen.
Negotiation requires judgment
A negotiation includes numbers, but it also includes timing, credibility, emotion, silence, internal politics, and the ability to understand what the other side really needs.
The best dealmakers know when a point is genuinely important and when it is being used as cover for something else. They know when to push, when to give, and when the relationship matters more than the immediate win.
A system can inform that judgment while the professional owns the decision.
The best model is intelligence plus judgment
The investment bank of the future should give senior professionals a level of market awareness that previously required a much larger team. It should continuously improve its understanding of companies, owners, buyers, lenders, transactions, structures, and intentions.
Then it should place that intelligence in the hands of people capable of turning it into action.
This is the model I believe in: technology that strengthens the banker and supports the client relationship.
Machines can expand what we see, and experienced people can decide what it means.
Create the market,
Jack