Blueprints for Success: Steering a Global Data & AI Architecture

Why do so many AI projects fail after the pilot? Merck shares its governance blueprint for steering proof-of-concepts all the way to production.

#1about 2 minutes

Understanding Merck's diverse global business sectors

Merck operates globally across life science, healthcare, and electronics, creating a complex environment for data and AI.

#2about 1 minute

The three pillars of Merck's data and AI strategy

The company's data and AI strategy is built on three pillars: fostering a data-driven culture, a hub-and-spoke operating model, and a central ecosystem called Optimize.

#3about 3 minutes

An overview of the Optimize data and AI ecosystem

The Optimize ecosystem integrates various data sources using a tech stack that includes AWS, Palantir, and Snowflake, offering services from dashboarding to generative AI.

#4about 2 minutes

Addressing the challenges of scaling a global data platform

With over 1,400 solutions in production, key challenges include aligning business sectors, preventing tool proliferation, and maintaining security at scale.

#5about 3 minutes

Managing new tool requests and controlling shadow IT

A formal IT tool request process and a "soft block" at the corporate firewall are used to guide employees toward approved solutions and manage shadow IT.

#6about 2 minutes

Guiding users with architecture documents and blueprints

Capability maps and architecture blueprints help teams select the correct tools within the Optimize ecosystem and follow best practices for implementation.

#7about 3 minutes

Implementing a mandatory architecture review process

A mandatory architecture review is required before provisioning AWS or Snowflake accounts and again before production deployment to ensure security and alignment.

#8about 2 minutes

Fostering collaboration with a hub-and-spoke model

Sector lead architects and monthly meetings facilitate communication and alignment, while the Scaled Agile Framework guides ecosystem development.

#9about 3 minutes

Using AI to enhance architecture governance in the future

Future plans include using AI to augment architecture reviews, create a helpful chatbot for developers, and automate the detection of unauthorized tools.