Ask a deal team where their timeline went, and the answer is rarely the model, the negotiation, or the lawyers' markups. It is the data room. Somewhere between the first index and the final report, three to six weeks disappear into document review — and most teams have stopped questioning why.
That is the diligence bottleneck. It deserves more scrutiny than it gets, because it is not one problem. It is four.
A mid-market transaction now routinely produces a data room of 2,000 to 10,000 documents: contracts, financials, board minutes, employment agreements, IP assignments, regulatory correspondence. The method for reviewing them has barely changed in twenty years — associates and analysts read, tag, and summarize, page by page.
The arithmetic is unforgiving. At a sustained pace of 40 to 60 documents per reviewer per day, a 5,000-document room is a multi-week commitment for a full team before anyone has synthesized a single finding. Volume grew tenfold; the method did not.
The costliest distortion is not the hours — it is who spends them. In most reviews, the people qualified to judge a change-of-control clause or a revenue recognition issue spend the bulk of their time on triage: sorting, deduplicating, locating the documents that matter. Judgment is the scarce resource, and the process spends it on filing.
Deal principals feel this as a quality problem disguised as a speed problem. When experts spend eighty percent of a review finding documents, findings get one read instead of three.
Diligence findings are worth the most early, when price and terms are still open. A material customer concentration issue surfaced in week one reshapes the negotiation. The same issue surfaced in week five, days before signing, produces a scramble — or worse, a post-close dispute.
Manual review inverts the value curve: the deepest findings arrive latest, precisely because reading is sequential. The bottleneck is not just slow; it delivers insight after the moment it could have changed the deal.
Review accuracy is not constant across a six-week engagement. Studies of professional document review consistently show error rates climbing as reviewers fatigue, and the documents reviewed in week five get materially less attention than those in week one. In a process that exists to find the one clause that changes the deal, declining attention is not a staffing inconvenience. It is diligence risk.
The instinct is to solve this with either more people or pure software. Both fall short in predictable ways.
More reviewers add coordination overhead and inconsistency — ten people apply a checklist ten ways. Pure AI tools read fast but cannot stand behind a finding; a probabilistic summary with no accountable reviewer does not survive an investment committee, and it should not.
The structure that works pairs the two deliberately:
This is the model we built LiquidDocs around: AI structures the data room in hours, and our analysts verify every finding and score the risk. Institutional-quality diligence, in days rather than weeks — not because the reading got faster, but because the experts finally get to spend their time on judgment.
For two decades, the diligence bottleneck was a fact of deal life. It no longer is. Teams that restructure the first pass are running reviews in a fraction of the time, with findings that arrive early enough to matter and a verification trail that holds up in committee.
The question for deal principals is no longer whether the bottleneck can be removed. It is how many more deals will absorb three lost weeks before it is.
If document review is where your deals lose their weeks, we should talk. Book a call — we will walk you through what a verified first pass on your next data room would look like.