Smarter Contracts, Faster Decisions: Legal in the AI Era ft. Gabriel Saunders

In this episode of "The Counsel's Code," we are joined by Gabriel Saunders, a leader in the legal tech and operations space. He shares his expert insights on the evolution of CLM from a simple tool to an AI-powered strategic partner, discussing implementation, AI, workflows, and the future of "agentic" legal tech.

Meet the Speaker: Gabriel Saunders

We are thrilled to feature Gabriel Saunders, a leader in legal technology and operations. He currently serves as the Director of Legal Operations at Exos, where he also sits on the Exos AI council, helping to shape the company's overall AI strategy.

Gabriel's unique career path began as an IT consultant to law firms, which led him to become an eDiscovery expert and build an early SaaS platform for government surveillance and unstructured evidence. After another startup in the food-tech industry, a general counsel introduced him to the world of legal operations, and he "fell in love with the less aggressive, more... warmer space of managing money and agreements."

Key Insights from the Conversation

From Tricycles to Teslas: The Evolution of CLM

Gabriel describes the evolution of contract management as a shift in "velocity and speed." The "tricycle" phase was managing contracts through "email... Slack messages... and dozens if not hundreds of folders"—a process filled with "wasted administrative time."

The "Tesla" is moving to a single, structured system that connects legal intake, drafting, e-signature, and storage. For him, this one change "literally took my job from 10 hours a day to four hours a day." This allowed him to stop focusing on administrative work and start concentrating on high-impact strategic initiatives, like the company's AI strategy.

CLM Implementation Nightmares: What Vendors Miss

According to Gabriel, most CLM implementations fail not because of the software, but because of "a level of expectation." The first mistake is a "set it and forget it" mentality. Software is not static; it's constantly evolving, and the implementation is never truly "done."

The second mistake is miscalculating the time required. "If you think something's going to take 3 months, just calculate it's going to take 6 months," he advises, to handle the "unknown unknowns." He notes that in his experience, implementation failure is almost always a "failure of will, not a failure of the software."

Balancing AI Automation with Human Control

When using large language models (LLMs) in contract work, Gabriel emphasizes the "human in the loop." He advocates for an "80/20 rule": "You should expect AI to do 80% of the work for you... and then you should expect to produce the 20% of refinement on your own."

He finds that most people who are disappointed with AI output "didn't understand how to give it context in the beginning" or they "talked to the same context window for too long" until the model lost context. The user, who understands the "subjective position of what is good," is the most important part of the process.

Workflows That Actually Work: The Kanban Approach

"I'm a big fan of Kanban boards. It runs my entire life," Gabriel says. As a visual person, he values how a Kanban board gives him an "immediate understanding of what your day is going to be like" in just three seconds.

More importantly, it allows him to see his entire team's workload at a glance. When a leader asks for a task "ASAP," he can instantly look at the board and set realistic expectations. "If I see that there's 50 to-do tasks on my five-person team, I'm going to try to mitigate those expectations."

Misjudging ROI: Where Legal Ops Go Wrong

The biggest miscalculation in ROI is underestimating the "time it takes to get adoption." Gabriel notes that while he got 60% of his company to adopt the CLM in three months, "to get like full adoption, maybe was a year and a half."

He achieved this through "polite persistence" and "white gloving" the experience for some users—making their intake tickets for them and sending them the links. He stresses the need for empathy, understanding that some users (like salespeople on the go) may have limitations, and it's legal ops' job to bridge that gap.

What You Stopped Doing: Prioritizing with AI

Gabriel's growth came from stopping "doing things that computers should do." This includes "cutting down the copy paste," moving folders, redundant storage, and sending repetitive follow-up messages.

"Computers are pattern recognition and pattern production machines," he says. "When you sort to produce a pattern, you should pass it off" to AI. The goal is to automate repetitive work to keep lawyers in a "flow state"—the sweet spot "right in between boredom and overwhelmed," where they can do what they went to law school for: "critically think and critically write."

Biggest AI Myth in Legal

The biggest myth is "that it's an easy button." People think they can just ask the AI to "write me a contract" and a "wonderful contract" will appear. Gabriel cautions that "working with AI is like learning any skill. It takes a learning curve." While working with a vendor can shorten that curve, it's still a new skill that requires user effort to master.

Proactive CLM: From Storage to Strategy

Gabriel notes that the "best part" of a modern CLM is the "thoughtful way... [it has] strung together all of the different point solutions." Instead of separate tools for drafting, version control, repository, and workflow, a true CLM combines them into a "single flow that an in-house team would need."

He also highlights a novel feature he sees emerging: "tracking the business obligations." This moves the CLM from a passive repository to a proactive tool that helps "maintain that the company is meeting those business obligations of the agreement."

What Should CLM Integrate With Next?

Gabriel advises CLM vendors "not to get caught on the side of fully structured data." He foresees a future "CLM in the background," powered by vector databases, that moves away from a standalone SaaS login.

In this future, the CLM "meets me in Word, you're meeting me in my email client, you're meeting me in my DMS." The functionality happens "behind the scenes," and the value is "just delivering to that to me through an AI experience."

Where Should CLM Go Next? ESG, Global Compliance?

Looking 5-10 years out, Gabriel sees a "transition from workflow to agent." A traditional workflow is a "very static" if-then command. An "agentic workflow" is more flexible, "allowing LLMs to make some more decisions."

This will enable "agent-to-agent communication," where a user can ask their CLM a question, and the CLM agent will "go ask my CRM agent a question about its database" and return a comprehensive, human-level answer. He sees this "agent orchestration" as the next frontier.

Your One Piece of Advice to CLM Vendors?

To achieve true enterprise adoption in large organizations, Gabriel's number one piece of advice is to "concentrate on really great integrations."

He notes that "the bigger the business you work with, the more in place the business systems are." A successful CLM must be flexible and able to "enter as a puzzle piece into a larger organization."

About "The Counsel's Code" Podcast

"The Counsel's Code" is your go-to podcast for exclusive interviews with top legal executives. Discover the strategies they've employed to cultivate their careers, excel in their positions, and emerge as true leaders in their organizations.

Throughout our discussions, we delve into the challenges of leadership and how these accomplished professionals manage the pressures that come with it. Our mission is to provide valuable insights and support for in-house counsel, fostering mutual growth and development.

Tune in for engaging and enlightening conversations with legal leaders who share their experiences, wisdom, and advice, creating a community where in-house counsel can thrive together. If you want to get featured, contact marketing@volody.com.

Disclaimer: The views expressed in this podcast are the speaker’s personal opinions and do not necessarily reflect the positions of the podcast, Volody, or any current or former employers.