How Aviva Uses Dataiku to Unlock Deeper Customer Insights

Use Cases & Projects, Dataiku Product, Scaling AI, Featured Catie Grasso

Understanding customer experience is essential in any industry, and multinational insurance company Aviva has long prioritized analyzing customer feedback to enhance user satisfaction. However, a key challenge arose: While categorizing customer comments provided valuable insights, it only painted part of the picture. Nearly 40% of feedback lacked detailed explanations, making it difficult to pinpoint technical issues with digital platforms like portals and apps.

A neutral or negative rating without context left Aviva’s teams guessing whether the issue stemmed from a technical failure, a service-related problem, or another unknown factor. To bridge this gap, Aviva sought a more comprehensive approach — one that combined customer feedback verbatim with data-driven insights.

Aviva logo

Turning Clickstream Data Into Actionable Insights

Recognizing the limitations of relying only on written feedback, Aviva saw an opportunity to integrate clickstream data — users’ interactions with the portal or app — to complement customer feedback. Clickstream data contains a wealth of information about user behavior and, by linking it with customer sentiment, Aviva aimed to build a holistic view of the digital experience.

This approach enabled the company to analyze how customers navigate its digital platforms, pinpoint where they encounter friction, and determine how those experiences align with their feedback. By incorporating application log events, Aviva could also provide technical teams with deeper diagnostics, accelerating issue resolution.

How Dataiku Helped Power the Transformation

To tackle this challenge, Aviva leveraged Dataiku’s advanced analytics and machine learning (ML) capabilities alongside Wipro’s data analytics and AI team. Since the team had already built a customer feedback categorization initiative on Dataiku, extending the solution to include clickstream data was a natural next step. Here’s how Dataiku made a difference:

  • Seamless Data Ingestion: The team ingested large volumes of clickstream data from Amazon S3 without friction.
  • Data Preparation & Preprocessing: Using Dataiku’s robust data preparation capabilities, Aviva cleaned and structured the data, grouping user sessions efficiently.
  • Automated Session Analysis: Dataiku’s visual recipes allowed for quick grouping based on user identifiers, timestamps, and key metrics, eliminating the need for manual scripts.
  • Mapping Customer Journeys: Pages accessed by users were mapped to specific journeys like claims or quotes, making it easier to analyze user behavior.
  • ML-Driven Scoring: A session scoring model, built using Dataiku’s ML features, evaluated customer interactions and pinpointed pages with high error rates.
  • Integrated Log Analysis: Dataiku’s ability to integrate with multiple data sources, including application logs, helped Aviva’s teams diagnose technical issues with greater precision.
  • Next-Level AI Capabilities: Plans are underway to leverage large language models (LLMs) to summarize log events into natural language, further streamlining troubleshooting.

Throughout the project, Dataiku’s user-friendly interface helped team members to collaborate. The visual pipelines, combined with the ability to use custom code where necessary, enabled us to achieve results more quickly — bringing us closer to a comprehensive view of our customers’ digital experiences.

-Mitesh Chandorkar, Data Science Architect, Aviva 

Driving Real-World Impact

With the Dataiku-powered analytics and ML solution now in its validation stage, early results suggest a significant impact on Aviva’s day-to-day operations. Here’s how:

  • A 360° View of Customer Journeys: The integration of telemetry data enhances visibility into customer interactions, providing a clearer understanding of where users struggle.
  • Improved IT Operations Efficiency: Linking customer feedback with real-time session data allows tech support teams to cross-reference logs and session details, reducing resolution times and improving customer satisfaction.
  • Data-Driven Decision Making: Insights from clickstream analysis help prioritize platform improvements, focusing efforts on high-impact areas such as frequently failing pages.
  • Actionable Management Insights: The solution not only identifies immediate issues but also provides historical trend data, helping teams proactively mitigate technical risks.

By combining clickstream data with customer feedback, we now have a much clearer view of where and why customers encounter difficulties, allowing us to resolve issues more effectively and proactively.

-Mitesh Chandorkar, Data Science Architect, Aviva 

The Value Unlocked With Dataiku

By leveraging Dataiku, Aviva accelerated time-to-value while empowering teams with data-driven insights. Some key benefits include:

  • Speed & Scalability: Dataiku processes large datasets efficiently, enabling seamless analysis of high-volume clickstream data without performance lags.
  • Enhanced Collaboration: The platform’s user-friendly interface allowed teams to work together seamlessly, combining visual workflows with custom coding where necessary.
  • Advanced Analytics & Automation: Built-in features like automated session scoring and time-series trend analysis streamlined the identification of patterns in customer behavior.
  • Future-Proofing with AI: As Aviva continues refining the solution, Dataiku’s AI and ML capabilities will enable even more sophisticated analysis, including predictive insights and automated troubleshooting recommendations.

The ease of sharing data insights through Dataiku’s platform leads to more informed decision-making across departments, from marketing and customer service to development teams, bringing more agility in process.

-Mitesh Chandorkar, Data Science Architect, Aviva 

Looking Ahead

With this new data-driven approach, Aviva has strengthened its ability to diagnose and resolve customer pain points. By integrating clickstream data with customer feedback, the company has shifted from reactive guesswork to proactive problem-solving — ensuring a more seamless digital experience for its users. As the model continues to evolve, Aviva remains well-positioned to harness AI-powered insights for even greater operational efficiency and customer satisfaction.

Overall, Dataiku not only accelerated our time to value but also empowered teams to work more efficiently and collaboratively. Its scalability ensures that as we continue to grow, we can seamlessly integrate more data sources and refine our solution further.

-Mitesh Chandorkar, Data Science Architect, Aviva

You May Also Like

How Aviva Uses Dataiku for Enhanced Scalability & Efficiency

Read More

The Evolution of NLP (& How Dataiku Can Help)

Read More

A Dizzying Year for Language Models: 2024 in Review

Read More