How to Close More Deals with AI in Investment Banking

We currently view AI in investment banking as a mechanism to increase productivity and efficiency. This is important but significantly misses on where we can increase revenue and dealmaking: leveraging AI to increase the probability of closing deals.

We can break down investment banking revenue streams into a simple formula:

# of deals × probability of closing deals × deal value

We will discuss leveraging AI to increase deal flow another time, but a crucial part of the equation that we often overlook is increasing the probability of closing deals at the best outcome possible. To illustrate how problematic this is, in 2025, approximately 31% of deals terminated without transacting. Currently, deal processes are not maximized for several reasons:

• The right buyers are not pitched correctly
• Key risks are not flagged early enough in a process where the management team and bankers can effectively mitigate
• They drag on

We can leverage AI to pull on three important levers to maximize the probability of closing at the best outcome possible:

• Buyer intelligence
• Custom materials tailored to each buyer
• Proactive risk flagging ahead of the due diligence process

Institutionalizing Buyer Behavior

We receive several signals regarding buyers from process after process. Partners and MDs typically know these signals intuitively, such as a buyer who traditionally underbids by 1–2 turns on EBITDA but reliably stretches at exclusivity or a buyer who consistently frames acquisitions around platform expansion and operational control rather than short-term financial arbitrage. The issue though is this is not institutionalized across the firm or maximally leveraged each time. No one will remember each individual interaction they had with a buyer, how they thought about a specific situation or concern six processes ago, and the list goes on.

As these signals and meta data stack up, we can leverage AI to ensure we are best positioning our client to the needs and behaviors of prospective buyers. The data we can leverage and analyze include:

• Prior deal behavior: what they have bought, how they have bid, where they have walked
• Strategic fit signals: adjacency, integration posture, declared priorities
• Behavioral signals: speed of engagement, question patterns, internal alignment indicators
• Underwriting tendencies: how they respond to concentration, cyclicality, complexity, and carve-out dynamics

The result? Firms can loop in the right buyers with perfect accuracy and create a methodical process that pulls in a data-oriented approach to relationship management. This creates the following understanding:

• Which buyers should be pushed early versus sequenced
• Which topics should be de-risked proactively for specific buyers
• What level of detail should be provided, and when
• Where senior banker time should be deployed to prevent predictable stalls

Buyer intelligence turns subjective relationship knowledge into structured process control. Instead of relying on memory and instinct, firms can design sequencing, messaging, and engagement around how each buyer actually thinks and underwrites. That precision directly increases competitive tension, reduces avoidable friction, and improves close probability.

Crafting Custom Materials

Catering to buyer preferences and behaviors does not stop at the underlying positioning and contextual understanding stage. We can customize all materials as well.

Given how intensive the process currently is, bankers create one teaser, one CIM, one management presentation, etc. Bankers will then send these same materials to a heterogenous pool of prospective buyers and investors, pitching the same narrative.

The problem is very clear: each buyer has different aspects that they look for. We can break this down to something as simple as strategics vs. sponsors but also be very granular and segment this down to the strategic acquirer looking to expand into an adjacent market vs. a sponsor who has a massive appetite for this asset and has lost in two processes in this industry earlier in the year. When buyers underwrite differently, a single “one-size” narrative creates friction. Systematic execution (target fit, diligence readiness, and the clarity of the value case) is repeatedly shown as the difference between deals that close and deals that stall.

At scale, heterogeneity requires heterogeneity – deal materials tailored to each specific buyer, touching on what will maximize their interest and intent to be more competitive in a process. Furthermore, there are different ways companies can fit into a portfolio or into a larger organization – this means that the framing and narrative to get there becomes exponentially more important.

Instead of sending the same CIM to 40 buyers, firms can deliver targeted narratives that speak directly to each buyer’s underwriting logic. This quantity and customization were not possible before. Now, AI directly enables this shifted paradigm at scale, where we can now emphasize the elements most likely to convert buyer interest into competitive intent.

Protecting Momentum in Due Diligence

Even if the first two pillars are acted upon, there are two other areas where processes can break down: slow timing and items caught during diligence that should have been resolved months before in a process. In essence, time kills all deals and late-process concerns about the company erodes confidence and momentum.

For the former, AI can ensure that all parts of the deal process can be accelerated, ranging from Q&A turnarounds and document retrieval to internal draft iterations and data room organization. For the latter, AI can flag key risks and where more support is needed at the beginning of a process. When vulnerabilities are understood early, deal teams can:

• Resolve inconsistencies before buyers see them
• Prepare mitigation narratives and data support
• Sequence disclosure in a controlled way
• Reduce the probability of reactive diligence spirals

Bankers already operate as proactively as they can. From running multiple processes, they know what to expect, which levers to pull, and how to ensure a deal stays on track. At scale, this becomes incredibly difficult to maintain, especially as deal documents and diligence questions pile up. Datasite’s anonymized data shows the average deal time on-platform increased from 6.9 months in 2021 to 10.2 months in 2024, underscoring how diligence has grown deeper and more complex.

By leveraging AI early on in a process, bankers can ensure they mitigate key risks before taking a company to market. They can also better prepare for how to ensure they are always best positioning their client and preserving negotiation leverage through disciplined process control.

Key Takeaways

Most failed transactions can be traced to three structural issues:

• Mis-targeting
• Misalignment with buyer narrative
• Surprise risk that destabilizes confidence

AI increases close probability when it reduces these systematically:

• Buyer intelligence reduces mis-targeting by improving process design
• Buyer-specific narrative modules reduce underwriting friction and increase competitive tension
• Proactive risk management reduces confidence collapse late in diligence and ensures tight processes

The investment banks that will win going forward will not simply draft faster.

They will use AI to design tighter processes, control buyer behavior, and preserve confidence throughout the lifecycle of a transaction.

That is how AI closes more deals.