In 2019, we invested in Arthur as the first production AI monitoring platform to detect and protect enterprises against modeling issues and biases. Fast forward just three years and the AI industry is moving faster than ever thanks to the rise of generative AI and innovations like ChatGPT. As quickly as the market has moved, Co-Founder and CEO Adam Wenchel and the Arthur team have worked to keep up with the changing needs of their growing enterprise customer base, which include some of the largest organizations in the world—including three of the top five U.S. banks, Humana, the Department of Defense (DoD), and many others.
We chatted with Adam to discuss his background in AI, how rapidly the market is moving these days, and his best advice for selling into the enterprise:
Arthur ensures customers’ AI systems are well-managed and deployed in a responsible manner. Last month, we launched a powerful addition to our suite of AI performance tools called Arthur Shield, which is the first firewall for large language models (LLMs). This patented new technology enables companies to tackle critical safety and performance issues in LLMs, like sensitive data leakage, “hallucinations,” prompt injections, and toxic language.
LLMs are one of the most disruptive technologies since the advent of the Internet. Yet, as with all new technologies, these advancements pose numerous potential risks to both companies and the public.
We’ve created the tools needed to deploy this technology more quickly and securely, so companies can stay ahead of their competitors without exposing their businesses or their customers to unnecessary risk.
I started working in AI in the ‘90s (!) as a researcher at DARPA right as I was finishing my CS degree at the University of Maryland. After a couple of years, I was lured into the startup world and haven’t looked back.
My previous startup, Anax Security, leveraged AI to detect and block cybersecurity attacks. After Anax’s acquisition by Capital One, I had the awesome opportunity to work with the CEO and CIO to start their AI division, The Center for Machine Learning, and scale it up to nearly 300 people as the VP of AI & Data Innovation. When I was deploying ML systems at Capital One to improve the systems being used in high-value areas like credit decisioning, cybersecurity, or fraud, there were no available solutions for ML monitoring and explainability—and it kept me awake at night. That’s why we started Arthur, and it’s been exciting to see how much it has resonated with ML practitioners everywhere.
With AI right now, since it’s changing so quickly, we found the best way is to lean into the role of providing thought leaders for open discussions with communities of AI practitioners so that we can all navigate this new world and make sure we’re all staying up to date on the latest AI developments together. This year, we started Ground Truth, an event series at our headquarters in New York City where we’ve hosted industry leaders to talk about topics ranging from generative models to data privacy to fairness and ethics in AI. We’re excited to host even more events and workshops this year to continue engaging with the community we’ve built.
Enterprise sales cycles are relatively measured in 2023, which was something we had to contend with, along with many of our peers. When ChatGPT launched in November 2022, I certainly could not have predicted the explosion of interest that it would create in LLMs, but I did see a huge opportunity. Our team has always operated on the cutting edge, so we were really well-positioned to take advantage of that moment.
The more we talked to people, the more we could see that LLMs were about to emerge as the top priority for enterprises.
We also realized that everything we’d been working on for the previous three years had changed radically overnight. We pulled together the team, said “we are tearing up the roadmap,” and spent the next few weeks talking to every person we could find who was building with LLMs to hear what walls they were running into. That led directly to Arthur Shield.
We knew we were onto something when we started getting calls from our customers saying “How soon can you let me start using it? I need it now!” The urgency people have felt about getting these models into production is truly astounding. Our team ended up rallying around the clock to deliver it two weeks ahead of the original schedule to help a few of those customers meet their own deadlines.
Our long-term vision is to make AI better for everyone. We want to make AI work better for the people who are deploying it, as well as for the people who it’s affecting.
We’ve been fortunate to have a lot of interest in each of our funding rounds. We definitely see big differences in the amount of value different investors add.
Work-Bench has been a great partner for us from the time we started our journey through our present growth. They’ve been with us every step of the way, and have been especially helpful in being a thought partner and helping us navigate enterprise customers.
Arthur is my second startup. The difference between my first and my second startup is that I actually worked at a major enterprise between the two, and it gave me an incredible amount of customer empathy. Enterprises run wildly differently than startups or tech companies, and if you take the time to really understand their unique dynamics, it makes it a lot easier to work with them because you understand what their motivations are and the pressures they’re under, as well as what makes them tick.
Startups aren’t an easy route, so make sure you choose a topic you are really passionate about. I’m fortunate that I’ve been in AI for over 20 years and I still get excited every day to learn new things and see where the world is going.
If you’re a Fortune 500 or tech company looking to access Arthur’s ML monitoring platform, contact the team here!
