Tel Aviv-based Superwise.ai today announced that it’s raised $4.5 million in seed funding, which it plans to use to scale its AI lifecycle management platform. This expanded platform, it says, will allow business and operational teams to take ownership of the health of AI models, giving them the ability to better trust the operation of their AI-driven processes.

“Business leaders are eager to implement and reap the benefits of AI, but quickly discover that complexity and lack of visibility creates serious trust issues once in production,” said CEO and cofounder Ofer Razon, who previously headed the AI business unit at Amdocs. “Most organizations don’t discover issues with their model until it’s too late and there’s already a negative business impact. This is partly caused by the fact that in most cases, it’s not even clear who within an organization owns the long term health and accuracy of AI models. Our team has years of experience dealing with this problem across dozens of enterprises, putting us in a unique position to guide our data-driven customers and give them trust in their models.”

Razon notes that inaccuracies can be introduced into AI models due to changes in the data that feeds them, market preferences, product updates, or marketing efforts. Errors by the model can result in lost profits, cybersecurity breaches, or, in the worst cases for industries like health care, even loss of life.

Superwise.ai

Above: The Superwise AI model monitoring dashboard.

Image Credit: Superwise.ai

Superwise’s platform, then — AI Assurance — leverages a performance management engine to provide monitoring and metric analysis for data science teams. It offers dozens of out-of-the-box KPIs (including several pertaining to bias detection, explainability, and AI analytics capabilities) tailored for commonly used models calculated in real time to measure statistics for models, with an AI-based component that automatically identifies issues before there are any business ramifications. Moreover, AI Assurance provides proactive actions and recommendations and executes corrective actions, leading to a claimed 25% reduction in data science team workloads.

VB Transform 2020 Online - July 15-17. Join leading AI executives: Register for the free livestream.

Superwise’s platform is akin to that of Fiddler, which taps an “explainable” engine that analyzes, validates, and manages AI solutions. Separately, companies including IBM, Facebook, Google, Microsoft, and Accenture have released tools designed to help spot — and rectify — bias in AI models.

Somewhat uniquely, however, AI Assurance integrates with a number of existing platforms, including AWS SageMaker, Google AI Platform, Azure Machine Learning, Kubeflow, H20.ai, and custom-built platforms. According to Razon, several undisclosed customers are already using it.

“We’re always attracted to innovative companies in new markets that are ripe for growth, and we see endless opportunity for Superwise.ai to continue growing thanks to its product maturity and experienced team,” said Capri Ventures managing partner Alex Pinchev. “The U.S. market is at the forefront of developing and implementing AI technology, and Superwise.ai already has the traction and customer success to put them miles ahead of the competition and in a position to become the market leader.”

The $4.5 million in seed funding was co-led by Capri Ventures and F2 Capital. It will be specifically put toward hiring talent to work closely with customers and open offices in both New York City and the Bay Area later this year, the company said.

Sign up for Funding Weekly to start your week with VB's top funding stories.