Breaking News

The Practice | Contracting Out

 “A contract is the most important data source in a company,” says Memme Onwudiwe. “It governs every relation with every employee, every vendor, every supplier, every customer. And that information is not just important to the legal department—it’s valuable to everyone.”

Onwudiwe is executive vice president of legal and business intelligence at Evisort, a contract management platform that leverages artificial intelligence (AI) to provide contract life cycle management and analytics to companies, legal teams, and anyone who deals with contracts. Onwudiwe, who graduated from Harvard Law School (HLS) in 2019, opted for a nontraditional career path in legal technology believing that he and his founding team members could transform contracting from an opaque administrative burden to something that could strengthen a company’s reputation and profit margin.

Indeed, as recent research from the Center on the Legal Profession (CLP) and EY Law indicates, contracting’s complexity often has financial repercussions—more than half of the organizations CLP and EY Law surveyed said that “inefficiencies in their contracting processes have cost them business.” While 92 percent of the organizations broadly agreed that contracting needs to change, exactly how they are trying operationalize that change remains uncertain. Moreover, while there is broad agreement on the need to do something about contracting, there is little agreement as to who “owns” the process—contracts switch hands and departments frequently, with legal, procurement, marketing, and others having their own goals and processes for the paperwork. This stymies both coordination and transformation, with 59 percent of those in legal departments believing they “own” contracting, while 39 percent of business development professionals believing they lead the charge. A “lack of alignment and clarity,” the report states, can have devastating consequences on the bottom line.

According to the EY-CLP Report, “While most understand the urgent need to change contracting practices, such transformation isn’t easy. Ninety-eight percent of organizations say they face critical barriers to delivering on their vision for contracting.” Above, a graphic from the report explains the top four barriers to transformation for organizations seeking to redo their contracting.

The CLP-EY Law study further highlighted the gap between technological adoption and usage. “Although most (70%) organizations have a formal contracting technology strategy in place, almost all (99%) say they do not have the data and technology needed to optimize their contracting process,” the report says. This means “many organizations face increased risk because they are unable to measure, manage and control adherence to their policy.”

Debates about contracting are often approached from the context of risk management and cost reduction. Indeed, in a separate CLP-EY Law report, risk management and cost reduction are two of the top five CEO priorities. In the first case, when it comes to contracting—and contracting technology specifically—software presents an opportunity to mitigate risk—for example, by reducing the opacity of a company’s contract portfolio. Noah Waisberg, CEO and cofounder of legal tech company Kira Systems, writes in a chapter on contract analytics in Legal Informatics:

Human-driven contract review has the conditions for a perfect storm of risk: by most accounts the work is achingly tedious, but at the same time serious consequences can result from errors. This calls for extreme attention to detail. Stories abound of missed provisions that were only found at the eleventh hour—or worse, after a deal was completed. The status quo contract review process is slow, costly, prone to human error, and can also generate initial results that are not as useful as they might appear. Contract review software can help change all that.

In the second case and with respect to cost reductions, as Onwudiwe notes, contracts are the lifeblood of an organization, being both a source of potential growth as well as a potential sinkhole. The relationship between cost reductions and contracting is clear. As CLP and EY Law reported, research from World Commerce & Contracting suggests “the average basic contract costs nearly $7,000 to create.” The report goes on:

Complex contracts, meanwhile, average $50,000. Given that large organizations, on average, manage 350 contracts per week, it’s understandable that 99% of organizations are planning to reduce the cost of contracting over the next two years. The scale of cost cutting being targeted is striking. At large organizations, slightly more than a third (34%) are targeting cost savings of 30% or more. Cutting one out of every three dollars from the contracting process cannot be achieved through small, targeted adjustments. Broader transformation will be required.

What should one make of this? Simply put, how the legal profession is evolving to deal with contracts is a prime example of legal informatics: using AI to understand, organize, and manage contracts means diving into, in Ron Dolin’s words, the “logic, structure, data, and measurement of law.” Innovation around contracts is also, as Dolin suggests in “Legal Informatics: Taking the tediousness out of law,” an important step forward for freeing up lawyers to tackle the work they enjoy, over the work they hate.

The legal tech landscape

Over the last few years, the legal tech landscape has exploded. According to September 2021 data from Crunchbase, venture capital has already invested over $1 billion across the legal tech industry just this year. Contract management solutions occupy a large piece of this market. In February, Evisort raised a $35 million Series B funding round, which was led by General Atlantic. Evisort’s clients include Microsoft (also an investor), Netflix, and Brooks Brothers. It sits among other contract companies, like Kira (recently acquired by larger legal tech company Litera with law firm clients like Davis Polk and Goodwin Procter), LawGeex (with clients such as eBay and Office Depot, as well as White & Case, which plans to deploy the software on a white label basis), and Ironclad (with clients like Mastercard and DoorDash). Market Research Future released a report in June 2021 forecasting that the contract management market alone will reach 6.5 billion by 2025. As Onwudiwe will tell you, however, there is contract management—which operates more like an Excel spreadsheet—and there’s contract intelligence, like Evisort.

In a webinar showcasing the technology, founding team member Riley Hawkins explains that Evisort’s algorithm combs through any contracts a client feeds it, extracting key data points such as dates and parties, termination rules, confidentiality, or force majeure clauses. You may be already recording this type of information manually, Hawkins says, but when done manually, the accuracy hovers around 60 percent. Evisort says that it does not release an algorithm unless it’s at 90 percent accuracy, and companies can help the AI learn so that they might achieve 100.

A man in a small corner of the screen shares his screen: on the screen is a contract with a sidebar extracting key information. At the top of the screen, headers like contract type, key provisions, dates, parties, are displayed, indicating that the technology can extract key information.

Riley Hawkins showcases Evisort’s customer-facing platform and how it organizes contracts. For full size, open the image in a new tab.

“When you’re turning a contract into data, you’re taking something subjective—like language and law—and turning it into something objective: numbers,” says Onwudiwe. Doing this, he believes, is a critical piece of being competitive in today’s marketplace. By providing a system that can standardize contract language, remind first-year associates who have inherited accounts to renew or sever contracts, or identify potential liability or error, Evisort allows businesses to learn more about where they could move from losing business to generating additional revenue. This, in turn, will transform the role of lawyers.

Where Evisort started and where it’s going

Looking back on Evisort’s early days, Onwudiwe reflects: “We didn’t really know what we wanted to build, but we did know what we wanted lawyering to be. And it was much less manual and it didn’t require data entry. It was less hierarchical and allowed folks to focus on the reasons they went to law school—doing complex legal analysis and not copy-pasting clauses.”

Onwudiwe was one of a team of six—four Harvard Law students, an MIT data scientist, and a computer scientist from Northeastern—who believed that they could make more impact on the legal profession by exploring the promises of AI than they could by following the traditional path of an associate at a law firm. Jerry Ting, who was a year ahead of Onwudiwe at Harvard Law School, first had the idea for Evisort (which stood for “Evidence Sort”) as an undergraduate. At the time, AI had been used for other sorts of tedious, time-consuming tasks, such as e-discovery, but not for building, managing, or reviewing contracts.

“What’s the point of a management system that’s just storing the document but not managing it?” asks Memme Onwudiwe.

In 2015, Ting joined forces with Amine Anoun, who was then a data scientist at Uber, and Jake Sussman, another HLS student. Onwudiwe met Ting and Sussman as a first-year law student while preparing for consulting company interviews. Both Ting and Sussman had spent their first-year summers at Boston Consulting Group, says Onwudiwe, and “not many people at Harvard Law School are looking to pursue nontraditional paths.” It was soon after that they invited Onwudiwe to join the company as head of sales, though, “At the end of the day, where there are five people at a company, everyone kind of does everything,” Onwudiwe says.

In 2016, the company took up residency in Harvard Innovation Labs, where, Onwudiwe jokes (with some seriousness), he spent more time physically than at the law school. There, over the course of three years, they built their proprietary algorithm. Onwudiwe relates those years as tedious but fruitful:

We thought it was stupid that you had to type all the stuff in the contract, and then type it all over again just to put it in a separate contract management system. It just didn’t make sense. You just typed those things in the contract, so what’s the point of a management system that’s just storing the document but not managing it? We thought it was ridiculous that we had to do all this manual entry of information and then do it again and again if you wanted any visibility into the contracts outside the data that was already tracked. Ironically that meant we had to do three years of manually tagging information people already had written. We did that to train numerous algorithms. The idea was that we’re tagging these documents so literally nobody will ever have to do this again.

Onwudiwe says that Evisort as a business has always given primacy to the accuracy of their technology. “When we started getting accurate algorithms, we were very excited, as you might imagine,” he explains. “And the lawyers said, ‘This is great. Let’s build a platform.’ Amine, our CTO, says, ‘Great, let’s find some software engineers.’ We said, ‘Wait, you’re not going to do this?’ He said, ‘I can’t do that. I can’t make a button bigger if your mouse goes over it. I can only make algorithms—I’m a data scientist.’”

This, Onwudiwe says, is something to look for when using a contracting platform: Is the company hiring data scientists? Prioritizing accurate AI, the company claims, distinguishes it among an increasingly competitive landscape. More recently, Ting wrote a column in Bloomberg Law warning that some companies who claim to use AI are in fact outsourcing review to humans. This, he writes, has ethical and legal ramifications that go beyond confidentiality, possibly breaching data residency and privacy laws, like Europe’s GDPR.

“You’ve got every past contract that you signed as a data point of where the market is and of where this negotiation will land. But if it’s unstructured data, it’s a burden,” says Onwudiwe.

“Outsourcing contract review is not new, and most legal teams go through a rigorous selection process when engaging a law firm or legal service provider,” Ting writes. “What is new is the explosion of automated contract review software with humans doing part of the review—sometimes to provide quality control, but other times stepping in for the AI when it falls short.” Whereas Onwudiwe wants law schools and firms to recognize how critical data literacy is for lawyers, Evisort aims to democratize the data, making both legal and non-legal jobs easier.

Putting the contracts into a system is only part of the process; to be useful, the data has to be accessible, particularly to those who need it when they want it. It also has to be structured and given context—in other words, it has to contain information. Right now, says Onwudiwe, if marketing wants to find out if they can put a logo on their website, they have to ask someone to sift through a 200-page document to find the single paragraph that contains the publicity clause. He continues: “If we could give that one paragraph to marketing in an accessible fashion in the first place, those are questions that never even have to go through legal, because that visibility and insight was available to those business stakeholders up front.” This, Onwudiwe says, is what will free the lawyer up to become a revenue-generating part of the organization.

Instead of just being the filter through which every contract question comes in now, lawyers can do more strategic thinking and actually say, “Hey, we’ve signed 100,000 agreements in the last 10 years. What have we agreed to?” If you think about it, that statement should be the most freeing thing, because that means any negotiation, you’ve done it thousands of times. So how could you lose it? You’ve got every past contract that you signed as a data point of where the market is and of where this negotiation will land. But if it’s unstructured data, it’s a burden. It is a folder full of scanned PDFs that keep you up at night. But if you can leverage AI to structure the data, then you’re coming into every negotiation with insight, with, “We talked to someone your exact size six years ago, and they landed here.”

The changing profession

While Evisort operated remotely for the last year and a half, Onwudiwe took some time to be a nomad around the country. Much of this he spent at home in Dayton, Ohio, where his mother, a doctor, serves as chief medical officer at Five Rivers Health Centers, a network of medical centers that prioritizes accessibility for low-income patients. Onwudiwe had already gotten much of the C-suite at Five Rivers Health Centers sold on Evisort: without an in-house legal team, the medical center was able to handle their own contracts seamlessly, including those pertaining to vendors and employees. One day, Onwudiwe’s mother saw him working at the kitchen table and asked, “Could Evisort work well for doctors’ licenses, too?” They didn’t know, so they tried. It worked.

In a year where medical staff did not have time to spare on finding and sorting through documents, Evisort was changing the game, both for large organizations and for regional health clinics in Ohio. Onwudiwe explains:

Because the medical centers are federally regulated, they have to deal with lots of audits, for instance, around doctors’ licenses. For example, the government will ask, are doctors actually licensed to do X kinds of procedures? But they also audit the process through which you keep track of that. What Five Rivers Health had before was a tool that was made specifically for this process. But that tool required manual entry and didn’t process the data it had. And even then it wouldn’t remind you that one is going to expire unless you actually type the date. There was no AI.

As Onwudiwe is the first to note, what Evisort offers is not revolutionary from a purely technological standpoint—it’s just revolutionary for lawyers. For years, Evisort had no MBAs, building out its employee base of product managers, salespeople, marketing, and more through lawyers—a testament to the creative and surprising ways that law school now trains students to think. Still, Onwudiwe says, law schools should be doing more to support entrepreneurship. This spring, Onwudiwe and Ting will return to HLS to teach a reading group, Startup Entrepreneurship and Innovations in Legal Technology, where they’ll prioritize voices of legal operations and nonlawyers to showcase what law can learn from other fields, as well as present options to students who want to do more than practice law in the legal profession. At a minimum, training other students to think of contracts as data might shift a new associate’s perspective—from language reviewer to data steward—and shift the profession toward reflective awareness of what measurement standards it has relied on for so long to determine quality.

In a June webinar CLP hosted with EY Law, Mary O’Carroll, chief community officer at contract management company Ironclad, echoed the sentiment that schools must change to accommodate the changing profession. With nontraditional roles emerging in legal spaces, such as data analysts, engineers, project managers, and more, she says, “Legal tech is emerging to support all this, but law schools and business schools are not starting to change their curriculum to adopt for all these new roles and the ways that lawyers need to think in the future—to try to get away from that traditional mindset to having an innovative one.”

At the end of the day, it’s all about data and time. Nonstandardized language costs lawyers time. Searching for the publicity clause costs lawyers time.

Beverly Rich, assistant professor at David Eccles School of Business, makes the point in Harvard Business Review that AI contracting will not reduce the need for lawyers: it will simply change how they use their expertise. As Onwudiwe has said, understanding contracts as data is critical for transforming the role of lawyers, allowing associates freedom from the rote tasks of sifting through contracts to find indemnity clauses and providing space for the critical thinking they spent years honing in law school. Now, Onwudiwe says, “lawyers can only show their value when they manage risky events.” With contract intelligence, however, they might be able to act with a preventive mindset. They can point out that “actually four percent of our contracts have a termination for convenience, allowing our customers to quit in 30 days. So, let’s go and renegotiate those, and that’s going to increase our bottom line. But you can’t have that conversation if you don’t have the data,” says Onwudiwe. Contract intelligence software will make it possible for lawyers to act as “counsel instead of contract reviewers,” writes Rich.

A recent story about Malbek, another contract management solutions company, notes that there are potential huge cost—and time—savings to be had. For instance, there may be hidden rebates in the contracts that are simply buried or favorable replicable terms from other contracts that new associates are forgetting to pick up but AI could help find—and help find quickly. Similarly, legal time can be cut by using AI to help identify whether lawyers or others actually need to deal with a particular contract or clause. At the end of the day, it’s all about data and time. Nonstandardized language costs lawyers time. Searching for the publicity clause costs lawyers time. Or, as Onwudiwe said recently on, The Convergence, a podcast with Harvard’s Negotiation and Mediation Clinical Program, when you’re scrambling to find all the force majeure clauses you’ve signed in March 2020 to find out if they mention “pandemic,” you need time on your side.