How AI Solved the Challenge of
Merger Review For an AmLaw 10 Firm



The Challenge of Merger Review

An antitrust counsel at an AmLaw 10 firm faced a problem that seemed unsolvable. She was looking for a better way to manage privilege review for merger reviews that trigger a request for additional information, or “second request.” A second request is a burdensome subpoena that requires an in-depth production of documents and data from the key players in a transaction. In the United States, mergers with a transaction value greater than $100 million must go through a pre-merger notification process. If the federal antitrust authorities (normally the FTC or DOJ) decide more information is needed or something about the transaction is worth investigating further, such as the effect of the merger on competition, the antitrust authorities can issue a second request.

In a traditional document review in response to a second request, the supervising attorney must manage hundreds of attorneys and quality control their work rather than focusing on the substantive issues in the case and the review. The counsel at this firm was introduced to this process as a young associate and found it to be “inefficient, expensive, and frankly inaccurate.” Human reviewers are capable of human error, and as she put it, “the people looking through the documents are not invested in the case.”

Exposing documents that might contain confidential or privileged information to hundreds of people also introduces its own inherent risk. With human reviewers comes the possibility of time-consuming false positives, unintentional disclosures, and, in the worst case, sensitive information being leaked to competitors or the media.

The counsel used technology as much as possible, including predictive coding for responsiveness, or relevance review. However, while predictive coding has proven to be effective for responsiveness review, it is not appropriate for privilege review. Privilege is not captured in the four corners of a document. Privilege depends on context, relationships between the people communicating, and the role those people play in the organization. Predictive coding is not capable of recognizing the larger environment of the document it is reviewing.

Even with a solution for responsiveness review, the counsel was still looking for a better way to handle privilege review and redaction. In her own words, “I was in search of a better solution for privilege; It was a major pain point in the process.”

“I was in search of a better solution for privilege;
It was a major pain point in the process.” 


Antitrust Second Requests


Improved Accuracy


Meet Compliance Deadlines


The Game Has Changed

To combat inefficiencies in the traditional document review process, the counsel decided to set up a proof of concept of Text IQ’s AI technology. Text IQ’s solution for privilege uses the proprietary Social Linguistic Hypergraph to understand the relationships between people, people’s role at a company beyond what is included in their job title, and company vernacular. Essentially, Text IQ’s AI is able to understand the contextual elements of a document that determine whether it is privileged without the need for a seed set or supervision.

Text IQ’s AI proved much more efficient and accurate than human reviewers. When asked about her introduction to Text IQ, the counsel said, “I told [Text IQ], if you can do what you say you can, this is gonna be game changing. We were able to pilot the technology, and they did what they said they could.” Consider the game changed.

Text IQ’s technology reduced risk by improving the accuracy and consistency of the review process and by limiting the number of people to whom each document is exposed. The AI technology works faster than human reviewers and is able to remove false positives, so less resources are wasted on a second pass review.

The counsel is now able to offer a better solution to her clients than the outdated, traditional approach involving a small army of contract attorney reviewers. Second requests are time-consuming, but time-sensitive subpoenas that have the potential to kill a major transaction. By utilizing Text IQ’s AI legal solutions, the benefits for the counsel have been threefold. She has upheld her legal oath to her clients by providing an efficient, accurate, and quality second request response, she has improved her firm’s reputation as an innovative and successful firm, and ultimately strengthened her clients’ relationship with the firm.

risk white


time white


cost white