April 09 2026
How Humans Should Deploy AI to Improve Deal Outcomes
The Fundamental Challenge
Most deal failures are blamed on cultural misalignment or rushed timelines, but the root cause often lies in how information is organized.
In engineering they call this the “context problem”. Systems fail when inputs are incomplete, unstructured, or disconnected. A machine learning model without proper feature engineering produces garbage. A software system without clear dependencies creates cascading failures. An API without documentation becomes unusable.
What “Deal Context” Actually Means
Context is not just “the document.” It is the network of facts, goals, assumptions, and history surrounding a request or finding - what matters, what was requested, what is missing, what is outdated, and how all this information relates. Consider what happens in a typical deal:
-
Legal reviews contracts in the VDR while finance analyzes spreadsheets in separate folders, neither aware of overlapping red flags
-
Commercial due diligence teams request customer concentration data without knowing operations already flagged supply chain dependencies with the same customers
-
Updated amendment documents arrive mid-process, invalidating or complicating earlier analyses that were never revisited
-
Key questions remain unanswered - even where Sellers responded, the Buyer teams often struggle to interpret how the question (and answer) matters to their deal strategy
Traditional diligence relies on disconnected tools: virtual data rooms or Dropbox for storage, Excel trackers for requests and progress, and email for deal-specific communication. This fragmentation creates version control issues, miscommunication, duplicate work, and silos where critical risks never reach the right teams.
Context is also lost for structural reasons: teams lack visibility into each other’s needs, engage content through limited function-specific perspectives, information degrades through handoffs, and critical knowledge lives in people’s heads rather than systems.
Why AI Makes the Context Problem Worse
As AI becomes embedded in diligence, context matters even more. Models don’t struggle with analysis as much as they struggle with missing intent and incomplete inputs. Bain’s survey of over 300 M&A practitioners found that many early AI users saved time generating outputs, but spent comparable time validating and contextualizing them.
The risk is that AI without proper context produces outputs that appear authoritative and complete, while built on flawed inputs. A polished diligence summary generated from scattered PDFs and partially-populated datarooms can create false confidence, leaving deal teams unaware of critical gaps or relevance to their investment thesis. In a compressed deal timeline, that’s arguably worse than no automation at all: it accelerates decisions rather than improving them.
Structuring relationships between documents, requests, findings, and deal rationale before analysis is what separates AI that surfaces insight from AI that automates dysfunction.
What Solving for Context Actually Means
Solving the context problem requires rethinking diligence as an information‑architecture and operating‑model problem. Successful firms are moving toward unified platforms that ensure:
-
Thesis Traceability: Diligence request scope should be shaped by the deal strategy, and every finding must be interpreted through the lens of the core investment rationale.
-
Continuous Context: When information is updated, the system should automatically signal which downstream analyses need to be revisited.
-
Cross-Functional Visibility: Findings across legal, financial, and operational streams must connect, diligence request scope should be shaped by the deal strategy, and every finding must be interpreted through the lens of the core investment rationale to create a single, holistic picture of thesis validation.
The companies getting this right are seeing tangible results. Organizations using end-to-end diligence platforms report 40–50% faster transaction timelines, 60% reduction in post-acquisition surprises, and 30–40% lower professional service fees. These benefits come not from AI alone, but from AI operating on systems that connect collection, verification, and analysis — preserving context from initial requests through closing and beyond.
The Path Forward
Billions of investment dollars go to virtual data rooms, analytics platforms, and due diligence services. Yet the core problem remains: firms optimize for review speed when the real constraint is context. The firms that will succeed in the next decade of dealmaking will be those that solve this problem - building systems that preserve coherence as complexity grows, surface connections as information accumulates, and ensure every stakeholder works from a shared understanding of the deal.
For practitioners evaluating their own readiness, try this test: at any point during diligence, can each workstream see how its findings trace to the underlying investment thesis? Can updated data room responses automatically surface dependency insights? When findings from Legal, Finance, and Operations converge on the same risk, can the system recognize the pattern - or does it require someone to manually connect the dots?
If your answer to these is “not yet,” the opportunity isn’t with faster extraction tools. It’s with a better way to capture, preserve, and use context - so humans and AI can drive better deal outcomes.
Noah Walters, Tower Software Labs | Gwen Pope, Tiger Team M&A
About the Authors
Noah Walters Noah is the founder and CSO of Tower Software Labs, an AI-powered due diligence platform backed by YCombinator. Tower helps deal teams centralize their work, collect and verify disclosures across disconnected sources, and extract insights from large volumes of unstructured data to spot issues earlier and close deals faster.
Gwen Pope is founder and CEO of Tiger Team M&A and veteran M&A leader with two decades in Fortune 500 M&A, with multiple patents for deal-process methodology. Tiger Team M&A delivers the first agentic platform for strategic decisioning in M&A, enabling serial acquirers to drive growth through M&A excellence.
