Customers do more with data through programmatic access for compliant ML workloads

An industry-leading data-as-a-service (DaaS) provider, operating globally and facing increasing demand for machine learning (ML) capabilities, partnered with Apheris to provide programmatic access to data while ensuring GDPR compliance. Apheris's non-movement of data and computational access governance led to several benefits, including quicker insights for their customers, reduced time and cost on regulatory compliance processes, and increased revenue by offering programmatic access for compliant ML as an added product feature.
Published 13 July 2023

Our customer: industry-leading data-as-a-service provider

With headquarters in the US and growing operations across Europe and beyond, our customer is a prominent data-as-a-service (DaaS) provider, serving multiple industries including finance, healthcare, insurance, and the public sector. Their on-demand dashboards and query interfaces enable their users to generate additional insights from unique datasets and related products. Responding to an increasing number of specific requests for granular data access, the DaaS provider is rising to the demand for increased ML and advanced analytics capabilities.

The Challenge: Navigating Regulatory Complexities to Meet Evolving Demands

Productizing data is an increasingly intricate task, given the diverse data privacy laws and industry-specific regulations associated with serving multiple sectors. With increasing business in Europe, adhering to the GDPR has emerged as a business-critical necessity. Moreover, the EU AI Act is poised to set new standards for the AI systems that their customers are developing. This shift not only ensures safety and respect for existing laws, fundamental rights, and values, but it also underscores the need for the data-as-a-service provider to be prepared to navigate these changes, as they directly impact the data assets they can commercialize and the way they service their customers.

As their customers delve deeper into machine learning and AI, the demand for unique dataset access and new use cases is escalating. Understanding which data can be utilized for specific purposes currently relies on manual processes that slow down sales and results in legal fees costing them tens of thousands. This considerable expense and inefficiency highlight the urgent need for an automated solution.

The Solution: Governed, secure, private programmatic access to data with Apheris

To address these challenges, the company piloted Apheris. For one of their industry use cases, they provided programmatic access to EU data to their global customer bases. This solution required no data movement, ensuring the company maintained full control over their data assets and satisfying a key part of GDPR compliance. Apheris' solution provided an additional layer to the company's data infrastructure, addressing several governance, security, and privacy challenges they were facing and easing the complexities around compliance.

The key features of the Apheris solution included:

  • Computational access governance: With Apheris' Compute Gateway, the company was able to control how their data was used by defining specific computations that could be run.

  • Asset policies: By defining who can do what, the company ensured that only permitted computations were run on the data.  This preserved privacy and IP, returning only anonymous results to the user.

  • Non-movement of data: Apheris ensured that data didn’t have to move, a critical aspect in meeting regulatory compliance requirements. Instead, computations are sent to the data meaning the company was able to offer new products to satisfy many of the ML use cases they were being asked for.

"Choosing Apheris allowed us to focus on delivering value to our customers without worrying about the regulatory complexities and privacy challenges associated with productizing our data."

Head of Data Products, Data-as-a-service-Provider

Results: Increased commercial opportunity, reduced risk

Adopting Apheris led to several notable benefits including:

  • Rapid insights: an 8-step process was reduced to 4, enabling customers to analyze insights and train ML models sooner. This typically means a 6-month process is now completed in 1 week.

  • GDPR compliance: risks associated with expansion in the EU were mitigated and customers are now able to train models on EU data compliantly.

  • Increased revenue: programmatic access for compliant ML is now offered as product functionality that has resulted in annual contracts of hundreds-of-thousands. They expect to roll this out to additional industry use cases and most of their enterprise customers in the next quarter.

  • Reduced costs and time: defining how data can be used once enabled scaling out to multiple customers, reducing the total time of the compliance process by 75%. This also resulted in removing thousands in costs associated with specific use case agreements.

Target state: Federated data sources are programmatically accessed by global users. Users securely send computations while Apheris ensures they are allowed and returns privacy-preserving results. All without centralizing data and the data custodian retaining control of their data assets.

"Implementing the Apheris solution has accelerated our customer onboarding process, making it twice as efficient compared to our previous operations. Moreover, we've managed to save thousands on legal consultancy expenses."

Head of Data Products, Data-as-a-Service Provider

Unlocking the value of sensitive data

This data-as-a-service provider shows how Apheris paves the way for businesses to unlock the full potential of their data assets in a secure, efficient, and compliant manner. Ready to unlock more value from your data?

Contact us

Machine learning & AI
Security
Privacy
Computational governance
Share resource to Linked InTwitter

Insights delivered to your inbox monthly