Project snapshot
The amount of data created worldwide exceeds 400 million terabytes a day, and IDC estimates that 83% of companies are actively investing in data initiatives. However, few companies understand how to translate their data into commercial value.
A global SaaS leader in the trust and identity verification industry wanted to find out. To support the launch of a new product that would fundamentally change their customers’ experience, the company needed to ensure their data would support a context-rich, automated analytics platform. The new product portal would allow customers to optimize the services they purchase from our client. For this service, a new modern data capability provides a single source for customer analytics and a new product. It also laid the foundation for AI.
While our client had already created a central platform with logic to generate analytics, the company had also gone through a series of mergers and acquisitions. As a result of their growth, they had a fragmented data environment with siloed data, from products that had been both built in-house and acquired through inorganic growth. To be successful, our client needed to democratize this data and break these silos.
During our initial assessment, Intellias found that the client’s environment had pockets of strong data capabilities, like advanced data science techniques and deeply embedded data governance policies. The challenge was to bring these capabilities together. Intellias identified the most commercially lucrative “backbone” use case that would underpin a roadmap for improvements. By releasing the analysis to their sales teams incrementally over the 9-month program, we were able to drive engagement and excitement for the platform. The new advanced data analytics capabilities have elevated trust in our customer’s business relationships, creating new sales opportunities and reducing churn. Working together, we successfully commercialized data into new customer pipelines and revenue streams.
Business challenge
Our customer had experienced a sustained period of growth that included mergers, acquisitions and reorganization. This led to the accumulation of large volumes of product and customer data spread across multiple teams. Each data set was stored and managed separately. Challenges included:
- Data compatibility gaps: Like trying to push a square peg through a round hole, the data in one system did not fit with the others. The lack of a unified, compatible data set prevented the company from comparing data across different systems.
- Slow time-to-value: For any analysis, the data had to be extracted manually, transformed, and then interpreted. This process delayed decision-making and restricted the value our client could create for their customers.
- Discovery lag time: Churn risk, reporting anomalies, and potentially valuable trends were not delivered in real-time. As a result, opportunities were being missed and our client was reacting to the data their products were generating instead of using it proactively.
- Customer intelligence gaps: The company lacked a consolidated view of multi-account customers, a consistent understanding of their customer’s risk tolerances, and single account or product customer views.
- Weak foundation for AI: The company’s existing data architecture and governance practices could not support scalable real-time analytics or advanced AI capabilities.
- Inconsistent data practices: The company did not have consistent policies for data capturing, data ownership, or quality control. This eroded confidence in the data set across the organization and had to be rectified manually.
These structural issues limited the company’s ability to move beyond verification services toward a deeper advisory role. Their customer-facing applications were operated without interacting with one another. Our client also lacked the ability to connect configuration changes to measurable outcomes. Furthermore, without a unified data platform, providing real-time, AI-powered analytics through a single portal would remain difficult.
Solution
Intellias partnered with AWS to design and implement a unified data platform for our client. A combined group of 30 AWS data platform engineers, data scientists, MLOps specialists, and our client’s product managers developed a phased delivery plan. The new platform replaced fragmented data systems with a single, scalable environment that could support automated analytics and customer-facing intelligence. Building a modern data platform on AWS became the core foundation for data ingestion, processing, output, and AI. The AWS data platform architecture also prepared the company for using AI in the cloud.
The solution included the following:
- “Backbone” use case: The key data use case to support the work was automated customer recommendations. It served as a guide to regular value and was released back to business stakeholders and sales teams while building a data platform on AWS.
- Central AWS platform: The unified platform was built using AWS services, including SageMaker Unified Studio for model training and monitoring, Glue and EMR for data processing, and Quick Suite for creating dashboards.
- Data ingestion: Separate identity platforms were integrated into a consolidated AWS data layer. While the platforms previously had independent architectures and data schemas, the unified architecture of the modern data platform on AWS ensured compatibility across different global regions.
- MLOps enhancement: Model training, versioning, monitoring, and retraining were formalized using AWS SageMaker. The cloud application replaced a manual workflow and provided a controlled MLOps lifecycle.
In addition to formalizing MLOps, the new AWS enterprise data platform aligned AI governance with the EU AI Act and older frameworks like GDPR to allow the processing of large volumes of sensitive PII data. Furthermore, it also moved the company from using experimental scripts to a managed ML environment on AWS.
Business outcomes
As a result of the AWS-based platform delivered by Intellias, the company was able to share performance data directly with their customers. They could see fraud trends, performance analysis, and context-rich configuration recommendations for their product portfolio. Additionally, the new AWS modern data platform provided many business values:
- The new data platform on AWS helped prevent customer churn through the identification of many suspicious transactions over 24 hours.
- An analysis of personally identifiable information (PII) enabled sales teams to strategically advise their customers. This improved customer satisfaction.
- The platform allowed the customer to open an entirely new revenue stream as part of their new product launch.
- It drew excitement, and demand generated by the sales teams throughout the program led to a backlog of new customers eager to access this insight.
- Their future roadmap now includes additional AI features, such as conversational AI that would allow customers to query data through the unified portal.
Technologies: Amazon Quick Suite, Amazon SageMaker Unified Studio, Scikit-learn, PostgreSQL, AWS S3, AWS Lambda, Custom RBAC, audit logging, Airflow, Python