Arteria AI, Goldman Sachs, And The Transformation Of Financial Legal Documents

1652812011 0x0

Much of the trade is made possible by legal agreements recorded in documents. This evidence has led AI companies to develop software that analyzes these documents and extracts important contractual provisions from them, which facilitates their evaluation and counting. Accounting (for auditing) and legal (for due diligence and commercial litigation) firms are rapidly adopting such tools.

Shelby Austin, an entrepreneur with a legal background, was aware of these trends and worked with companies on them as an attorney at a large law firm and at a legal services startup. When her startup was acquired by Deloitte, she observed the challenges of extracting contractual data from clients. When she became the head of Omnia, Deloitte’s AI business in Canada (disclosure: I’m a senior consultant for Deloitte’s US AI practice, but I don’t work with the Canadian company), she saw how AI was beginning to be used for documentation processes. . But something bothered her: why go through a traditional process of creating a contract in the first place, then extracting the salient details later? She and a few colleagues, including fellow lawyer Abrar Huq, started working four years ago on an approach to improve the documentation development process with AI by creating structured data from the start. This company was spun off from Deloitte in 2020 and became Arteria AI. Austin is the CEO and Huq the chief revenue officer.

Huq described Arteria’s goal with its customers: “Once we have the data and the trust of the user, we can automate the document creation and review process. The vision is twofold: First, to enable our customers to digitize their contracts and feed downstream systems quickly and accurately. Second, access predictive analytics and machine learning to reduce friction in trading processes. We want to take the knowledge embedded in client trading manuals (which may be a few hundred pages long and may not even be used) and enhance it by analyzing data. We can then integrate this knowledge into a more automated trading workflow. This would reduce both the cycle time and the level of risk.

Arteria’s business model

Arteria, which is now deployed for customers on 4 continents, is part of the growing segment of “contract lifecycle management” or CLM software. It focuses in particular on the “dematerialization at source” of contractual documents within financial services companies. He helps his clients create structured data when negotiating with their clients and creating contracts. Arteria attempts to capture data early in the process, rather than at the end (or not at all), and manage the workflow and data as it travels down the contracting process.

It’s not that financial companies create documents out of thin air, but the situation doesn’t get much better than that. They usually have SharePoint libraries of unstructured contract templates in Word or PDF format, for example, important documents like a “credit support schedule”. The problem is that they usually have hundreds of different versions of the same document in their libraries, each one missing metadata. They should browse the alternate templates and modify them to meet the needs of their specific offering, creating another unstructured document for the library.

Arteria takes existing unstructured documents, analyzes them and extracts their key contractual elements into structured data. Key data elements (currency, for example) can be aligned to a data model or taxonomy. Most sophisticated financial services companies already have an existing data model, typically used for ex post facto data mining, which is loaded into Arteria.

The most important idea, however, is that new documents are created with structured data from the start. Arteria guides document creators through a workflow that assembles key data elements and data values ​​into contract documents. The workflow can be personalized for each user: creator, negotiator, approver, etc. The life cycle of the contract would take place not only within the issuing company, but also within the counterparties with which the negotiations take place. The eventual vision is that AI would be used to help both parties negotiate, interpret counterparty markup language, and quickly build consensus among all parties to a contract. Such a smart, semi-automated process could eventually reduce the demand for human lawyers in contract negotiations, but no one seems to believe that would happen anytime soon or that many lawyers would lose their jobs.

A new activity that lawyers involved in contract negotiations could devote their abilities to is the analysis of data and statistics that will emerge once contracts are composed of structured data elements. With enough data to train machine learning models, contractual terms that minimize risk to the originating party and maximize the likelihood of acceptance by counterparties could be automatically inserted into contracts. Lawyers could be informed by a system that “this provision has been rejected in 80% of contracts; reconsider its inclusion? Eventually, a much smarter and automated process for creating and modifying contracts could emerge. However, this cannot be done without the use of structured data in the trading process.

Arteria and intelligent document automation at Goldman Sachs

Goldman Sachs, one of the world’s largest investment and financial services companies, makes extensive use of contract documents in its business dealings with its counterparties. The company is sufficiently focused on contract documents that there is a global head of policy and documentation in its Global Markets unit – Tuvia Borok, a managing director based in London. Borok has over a hundred people in his organization and he is leading the effort to create more intelligence and automation in contracting processes on the business side. So far, he has been running the project for 18 months. His counterpart on the technical side is Mike Pieck, a managing director in New York who is a Technology Fellow.

Borok and Pieck described the document creation and negotiation processes. “It’s slow, very manual, very unstructured. It can take months for these documents to be executed. Borok added, “Everything in the legal sphere is a manual process. We are all eager these days for information, but in the contract phase things are slow with little accessibility.

Goldman is using Arteria and other software tools to radically rethink the contract lifecycle. Borok said the main goal is to bring some leverage to the process and improve the employee experience. “We have highly qualified and trained legal professionals,” he noted. “We would like to use their skills more effectively. Searching through other document templates and modifying them is not a good use of their time or abilities. Pieck added that the redesign effort is also focused on extracting value from contract data. “There is a lot of value in the resulting documents and facts that need to be monitored. We cannot access it in unstructured documents. We need to understand what they contain and operationalize the terms.

The plan to change the way contracts are created and negotiated is as much about organizational and cultural change as it is about technology, the two Goldman executives argued. Borok, who himself was a practicing lawyer for 17 years, said many in the legal profession worry about the impact of automation on their work. He too was scared at first, he says. The legal profession, he commented, is not data-centric or automation-centric. The cultural acceptance of such abilities isn’t there yet, he said, and that means a lot of change.

Sometimes he tells his team, “Think like a fintech,” which broadens their horizons. He also emphasizes to his team that AI and automation are only meant to make the job of contract negotiators easier, not to replace them.

Pieck commented: “These changes will take a long time. We need to change not just Goldman Sachs, but the entire legal and investment infrastructure. The only way to do that is to establish small moments of credibility along the way.

Source link

READ MORE:  6 client management tips for advisors

Leave a Comment