Manually processing legal contracts can lead to great inefficiencies, but it doesn’t have to be that way, says Troy Pospisil, founder and CEO of San Francisco-based Ontra. He argues that information technology professionals at law firms can help bring costs down significantly with new AI tools.
Pospisil spoke with StrategicCIO360 about just how much manual processing can cost, why automating contracts is like semi-autonomous driving and how AI is affecting talent pressures at law firms.
What are the drawbacks to manually processing legal contracts, and how can IT leaders help improve their firms’ efficiencies?
When you consider the multitude of contracts firms need to process every day and how long a single contract takes, you realize how much of a time sink contract creation is. A recent survey found 27 percent of business and legal professionals reported taking between one and three full days to turn around a contract—and 10 percent reported a typical four-to-seven-day timeframe for a single contract.
The process of implementing contract automation pushes firms to determine preferred and fallback terms, resulting in more consistent language across contract types. Then, automation can offer a more efficient and faster process based on these predetermined contract terms, AI suggestions and streamlined workflows. Additionally, insourcing routine contracts costs more than firms initially believe.
For example, a medium-sized financial services company in New York mistakenly thought they spent $133,000 per year on processing NDAs in-house. However, after a thorough calculation that included direct and indirect costs of employment, lost productivity, employee churn and outside counsel expenses, the total cost of insourcing was $1,145,000 annually.
How can IT professionals at law firms ensure contracts are processed accurately with automation technology, given the breadth of different contracts and legal agreements, especially in a vertical where data needs to be 100 percent accurate?
Even routine contracts pose difficulties that AI can’t address on its own yet. There are too many nuances to the law and how parties may want to navigate a transaction. A “human-in-the-loop” model combines artificial intelligence and human insight. After the HITL model predicts contract terms, a legal professional either confirms the prediction or rejects it and offers an alternative.
The result is a stream of feedback that trains the model further and improves the accuracy of the AI’s suggestions. Ultimately, HITL contract automation produces a better—and faster—outcome than either lawyers or AI could manage independently.
You make a great point regarding the critical need for accurate data, which is why we see HITL solutions as the best way to deploy contract automation for legal teams. Human oversight within contract automation ensures that over time, the AI predictions for routine contracts become even more accurate, and parties can negotiate less.
With all of this data on contracts processed, at some point can the industry see a more standardized contract or template of sorts?
When lay people think about a contract such as an NDA, they might assume these documents are already uniform and standardized, just long and filled with legalese. But every agreement needs to factor in the unique circumstances of the transaction. Reading that, you might ask yourself, “how can there possibly be a contract template if each case is unique?” That’s where HITL technology comes into play.
Take self-driving cars, for example. No manufacturer could roll out fully autonomous cars instantly. First, manufacturers started collecting data. Year after year, they’ve analyzed their data and improved their technology—now semi-autonomous driving technology exists.
Semi-autonomous vehicles and the legal industry are similar in that both are complex. Every road and driver are different, just like every transaction is unique. But an immense volume of data related to vehicles, driver responses and maps exists. Similarly, we have an immense amount of information regarding the terms included in NDAs and other routine contracts. Through AI-assisted contracting, lawyers can quickly negotiate and finalize routine contracts, no matter who introduces the initial draft.
The HITL technology respects that routine contracts will have a large swath of language in common while still requiring unique adjustments. The solution can offer standardized language while allowing lawyers to make necessary changes based on the parties’ specific preferences.
What impact will the role of contract automation have on jobs in the future?
Whenever anyone throws around the word “automation,” some assume robots are taking away jobs from humans. This could not be further from the truth for contract automation in the legal industry.
Typically, firms save tasks related to high-volume routine contracts for junior associates and analysts straight out of business school. While everyone needs to start somewhere, when firms automate routine contract processes, these team members can use their talents for more thought-intensive, strategic projects instead. Not only does that make these employees feel more valued and like they’re adding value to the company, it also supports their professional development, improves work-life balance and lowers the risk of burnout.
Firms have to prioritize all these factors as the Great Resignation hits the legal industry. Leopold Solutions, a New York-based legal market research company, reported the average associate attrition rate in Am Law 100 firms was 27 percent in 2021. So, instead of robots taking over law jobs, they’re improving existing roles by giving associates more meaningful tasks.