Rich Robbins Believes Slow and Steady Wins the Innovation Race

In our Innovation Spotlight series, we interview innovation leaders about how they approach their jobs and how they’ve implemented and deployed Harvey.

In this edition, we chat with Rich Robbins, Director of Applied Artificial Intelligence at Reed Smith.

Rich previously served as a partner at a major global law firm where he created the knowledge management department, General Counsel for Morningstar, Inc, and Managing Director of Applied Artificial Intelligence at Epiq Legal Solutions. He holds bachelor’s and master’s degrees in electrical engineering and computer science from the Massachusetts Institute of Technology, a juris doctor from The University of Chicago Law School, and a master of information and data science from UC Berkeley, where he helped create its graduate-level class on generative artificial intelligence.

At its core, innovation is about change. It is a means to achieve things in new ways, some dramatic and often mundane. What is innovative to me may be old hat to you, and that’s okay.

Innovation requires a willingness to challenge the status quo, but also restraint. I do not believe in change for the sake of change, and sometimes the status quo is that way for a good reason. But our eyes must always be open to the possibility of change. We cannot see what we cannot imagine.

I delight in my family — I hope that brings me a better perspective to all that I do.

I am a lifelong technologist. At my core, I am an engineer and computer scientist with an academic bent. I love to teach. I enjoy writing and communicating, with precision and nuance. To teach something you need to understand it — really understand it.

At work, I am always teaching. I consider myself successful when someone I have taught is now able to teach the topic to others, preferably in a room where I can sit in the back and smile.

I have been wrestling with information and language-rich knowledge problems for roughly 45 years, 35 in the law.

Today, our profession is getting closer to creating the kinds of tools we’ve always wanted. I have a reasonable intuition when it comes to identifying trends or capabilities that will matter in the future. I know what it feels like to see something on the horizon and get excited and then share that excitement with others as things shake out.

Reed Smith is not shy about new technology. We are eager to explore what’s possible well before it becomes the norm. So Harvey’s potential resonated with the firm early on. With Harvey as our primary AI platform, we give our attorneys and everyone else at the firm an incredibly potent and ever-evolving tool that’s built for our work.

I have worked on technology adoption projects at law firms for most of my professional life — as a technology-savvy attorney, as a senior member of firm professional staff, and as a consultant. What we have seen at Reed Smith with Harvey is by far the most successful firm technology adoption story I have ever been a part of.

We moved from pilot to soft launch in May 2024. That month, approximately 80 firm employees used Harvey. A month later, that number had reached about 200. By May of this year, we were at about 800 monthly users, and now we see weekly user counts around 500.

There has been meaningful adoption across practices, regions, and seniority levels. We expect all of our associates to complete Harvey training and be confident with the solution, and are just shy of reaching that goal.

There was an initial “this is for the associates” or “give it to an internal service desk” attitude. What has surprised and delighted me is that an increasing number of partners have embraced Harvey and seen positive results. This has resulted in a growing community of internal senior evangelists, and in that role we are currently at about one-third adoption.

Slow and steady wins the race. High-profile, one-time events like a massive hackathon, a special training lunch, or an internal competition can too often be like a heavy rain in the desert — here today, gone tomorrow.

Therefore, we take a persistent, multi-pronged approach. Adoption is not a one-time thing. We use every internal firm communication channel available to us:

And, of course, I have given many presentations. So many presentations. All over the world.

Our senior leaders set the tone for the rest of the firm. They amplify our message and have become among our best advocates. Now that we have momentum, the program seems to have taken on a life of its own. I’m spending more time training others to deliver the message instead training on Harvey itself.

I will start with something a bit meta. I love showing people how to use Harvey to write better prompts for Harvey.

That’s a skill that may become less important over time as Harvey and other tools of its kind get better at enriching prompts but simply suggesting to people that they can use Harvey to craft better prompts for Harvey is powerful.

One of our power users has become an artist at blending his use of Harvey with a traditional eDiscovery platform. He moves back and forth between the two to use each tool to its best. People don’t always think about how to make effective use of their tools in concert. Sometimes you need both a hammer and a screwdriver.

Another power user is a business development professional. He leads his team by demonstrating how to use Harvey alongside background materials and past work to craft more personalized, client- or prospect-specific communications.

For repeatable tasks where we have reference data for the “traditional” way of doing things, our goal is to produce the same or better-quality work, faster.

When it comes to GenAI, it is more about whether the tools help us expand scope. Can we do things we simply couldn’t do before, without undue burden or cost? For example, GenAI helps us analyze a large collection of documents for inconsistent or shifting answers quickly on very short notice.

For now, we use engagement as a proxy for success, assuming that people will only use Harvey if they find it valuable. We track the number of unique monthly users, and how many queries those users made. We also conducted an internal Net Promoter Score-style survey with terrific results, which we plan to repeat periodically.

First, setting reasonable expectations is essential.

Second, GenAI is not a spectator sport. People must engage. Lectures and cheat sheets are not enough. Enforce the notion that the tools will not turn novices into experts — those who are most proficient at a task will see the best results.

I believe it will allow attorneys to deliver greater value, more efficiently.

That will impact what work gets done where. Tasks typically outsourced to firms, or perhaps ALSP, may be brought in-house. Perhaps that will create a democratization of which firms are well-placed to do what work. Perhaps instead of reducing the need for attorneys, this will expand what attorneys can achieve because we are able to attack problems that we could not before.

After all, whether you think the quote comes from Yogi Berra or Neils Bohr: "It's tough to make predictions, especially about the future."