arthur.ai Raises $15M Series A for ML Monitoring

With an increasing number of organizations developing machine learning models, Arthur.ai aims to address the challenge of maintaining model accuracy as performance can decline over time, impacting their intended precision. Reflecting growing demand for solutions of this kind, the company today announced a $15 million Series A funding round, even amidst the ongoing pandemic.

Index Ventures spearheaded the investment, with participation from new investors Acrew and Plexo Capital, alongside existing backers Homebrew, AME Ventures, and Work-Bench. This new funding arrives approximately one year following the startup’s $3.3 million seed round.

According to CEO and co-founder Adam Wenchel, data scientists typically build and rigorously test machine learning models in controlled laboratory settings. However, once deployed into real-world applications, these models can experience performance degradation. Arthur.ai is engineered to identify and address these instances of declining performance.

Despite the widespread disruption caused by COVID-19 this year, the company has experienced a 300% revenue increase over the past six months. “Throughout 2020, we’ve expanded our outreach and engaged with a growing number of customers, resulting in significant initial traction with both established enterprises and innovative tech companies,” Wenchel explained. Currently serving 15 customers, the company is observing strong resonance with its solution.

Notably, Amazon Web Services recently unveiled a comparable tool, SageMaker Clarify, during their re:Invent event. However, Wenchel views this as confirmation of the market need his startup is addressing, rather than a competitive threat. “This announcement helps raise awareness, and given our exclusive focus, our capabilities extend beyond those offered by major cloud providers,” he stated.

Mike Volpi, an investor at Index, recognizes the inherent value of this company. “A crucial component of the AI infrastructure lies in performance monitoring and risk reduction. Essentially, is the AI system functioning as expected?” he noted in a blog post announcing the funding.

A year ago, the company consisted of eight employees. Today, the team has grown to 17, with plans to double in size again by the end of the coming year. Wenchel emphasizes the importance of a diverse workforce, particularly for a product focused on identifying various forms of bias. He highlights the significance of diversity within the investment team and board, which he has successfully achieved, and extending this principle throughout the organization.

“We actively sponsor and collaborate with organizations dedicated to coding education for underrepresented groups, including those specifically focused on AI. We intend to continue this support, and once in-person events resume, we plan to actively support organizations like AI for All and Black Girls Code,” he said. He believes these partnerships will establish a pipeline of talent from underrepresented communities, enabling the startup to meet its future hiring needs.

Wenchel expressed his desire to return to a partial in-office work model when feasible, believing that certain collaborative tasks would benefit from in-person interaction.