Arthur.ai, a New York-based startup that helps companies monitor the accuracy of their artificial intelligence applications and find areas for improvement, has secured $42 million in new funding.

The company announced the Series B funding round today. The round was jointly led by Acrew Capital and Greycroft. Arthur.ai, which is incorporated as Arthurai Inc., previously secured $15 million through a Series A investment that it announced in 2020.

Large-scale AI models can often become less accurate over time. One common factor behind accuracy declines is a phenomenon known as data drift. The phenomenon occurs when an AI is configured to process a certain type of data during development, but is given different data to process once it’s deployed in production.

Arthur.ai has built a software platform that can automatically detect data drift. The platform is also capable of detecting fairness issues in AI applications. It visualizes issues in a dashboard that engineers can consult to identify the scope of an AI malfunction, as well as find its cause.

The company has equipped its platform with an alerting tool that can notify developers immediately after a machine learning application experiences a decline in accuracy. Additionally, the platform makes it possible to compare the accuracy of different neural networks. Developers can test multiple versions of the same neural network and deploy the one that proves most reliable.

According to Arthur.ai, its platform is implemented as a collection of microservices optimized for scalability. The microservices-based architecture allows the platform to process up to 1 million data operations per second. As a result, it can be used to monitor the accuracy of large-scale AI models that ingest a significant amount of data.

Companies can deploy the platform both in the public cloud and on their on-premises infrastructure. According to the startup, built-in cybersecurity features make it possible to limit user access to the platform’s core components and the data that it collects about a company’s AI applications. 

Arthur.ai says its latest funding round follows a year in which it grew annual recurring revenues by an average of 58% per quarter. Overall, it has more than quadrupled revenue during the past 12 months. In a bid to continue its recent sales momentum, it will use the proceeds from the funding round to grow its AI research, customer success and data science teams.

The company operates in a competitive market. Arize AI Inc., a startup with a platform for troubleshooting technical issues in AI software, closed a $38 million funding round earlier this month. Another market player, Diveplane Corp., raised $25 million a few days later. 

Photo: Arthur.ai