ZS Associates deployed AI agents in Dataiku to speed patient-journey insights, automate root cause analysis, and keep 100+ ML models running reliably — reducing delays, operational risk, and manual effort at scale.
60%
faster root cause analysis turnaround
Millions
of dollars in annual cost savings
25%
reduction in post-deployment errors
The Challenge
ZS Associates is a global management consulting and technology firm that helps organizations improve performance by turning data, science, and technology into better business outcomes. Across healthcare, financial services, and consumer industries, ZS relies heavily on advanced analytics and machine learning (ML) to support high-impact patient, clinical, and commercial decisions.
As data volumes grew and ML systems scaled, however, teams began to encounter friction in two key areas: patient-journey analytics and MLOps. Analysts struggled to quickly find and connect insights buried across hundreds of decks, PDFs, and reports, slowing time-sensitive decisions and increasing the risk of missed signals.
At the same time, ZS was running more than 100 production ML models, each producing complex and inconsistent error logs. Diagnosing failures required fully manual root cause analysis, with engineers combing through logs line by line and documenting issues under strict SLAs. This reliance on a small group of subject matter experts (SMEs) led to delays, avoidable downtime, and added operational risk. Together, these challenges made it harder to move quickly, maintain model reliability, and scale AI without increasing cost or complexity.
The Vision
ZS Associates set out to improve performance across both analytics and operations without adding new tools or headcount. They aimed to streamline knowledge retrieval, improve model reliability, and create a modular, repeatable architecture for future GenAI projects, so new use cases could launch quickly without rebuilding workflows or governance from scratch.
The Execution
Using Dataiku, The Universal AI Platform™, ZS Associates built two AI agents on a shared, reusable foundation, allowing different teams to solve distinct problems using the same operational approach.
1. Patient-Journey Knowledge Agent
ZS Associates built an AI agent that allows analysts to retrieve patient-journey evidence in seconds instead of hours, eliminating manual searches across decks, PDFs, and unstructured files. Using RAG, the agent scans the content so teams can ask natural-language questions and surface the right information instantly. The result is faster, more consistent insights, less time spent searching for information, and a standardized approach that can be extended to new therapeutic areas with minimal effort
2. Automated Error-Log Analyzer
For MLOps, ZS Associates built an AI agent that automatically analyzes error logs across production models, identifies likely root causes, and generates summaries in minutes instead of hours — reducing work hours and reliance on a small group of SMEs, improving model uptime, and delivering millions in annual cost savings.
Together, these agents enabled ZS Associates to standardize AI workflows across analytics and operations to accelerate insights, improve reliability, and reduce manual effort without increasing operational risk.
For the future, this use case is not just a one-off solution — it has established a blueprint for scaling GenAI responsibly within our organization.
— Dhwani Shah, Associate Consultant, ZS Associates
What’s ahead for ZS Associates? The company is extending reusable AI patterns into new domains to speed time to market. This makes it faster to launch future GenAI initiatives while maintaining consistent governance, reliability, and performance across the organization.
GET TO KNOW ZS ASSOCIATES
INDUSTRY
HEALTHCARE & LIFE SCIENCES
TECH STACK
Dataiku + Snowflake + AWS + Microsoft Azure + OpenAI + Anthropic
KEY CAPABILITIES
GenAI & Agents, Data Prep for AI, AI Governance