Data Governance in the Era of AI

An $18,000/month data warehouse confidently delivered wrong answers. The problem wasn't the AI, but the data.

#1about 5 minutes

The high cost and failure of ungoverned data

An initial strategy of collecting massive amounts of data led to high costs, complex queries, and AI models that produced incorrect results.

#2about 2 minutes

Establishing clear data ownership with data mesh

The data mesh philosophy makes data producers responsible for their data, creating clear ownership from domain systems to domain datasets.

#3about 2 minutes

Defining the six essential aspects of a data product

A true data product must have six key qualities: ownership, documentation, data quality checks, architecture, a data contract, and security.

#4about 2 minutes

Prioritizing data governance with a tier-based system

Classifying data products into three tiers based on business impact helps focus governance efforts on the most critical assets first.

#5about 1 minute

Using data contracts to manage critical tier one data

Data contracts are formal agreements between data producers and consumers that define responsibilities, SLAs, and procedures for handling issues with critical data.

#6about 1 minute

Overcoming the cultural challenge of data governance

Implementing data governance is not just a technical fix but a long-term cultural shift that requires changing the entire company's mindset around data.

#7about 2 minutes

Measuring the impact of data governance initiatives

Implementing governance dramatically reduced the number of tracked interactions, queries, and dashboards while improving data completeness and analyst productivity.

#8about 2 minutes

Enabling successful AI with a governed data foundation

With a governed data foundation in place, AI can be successfully used for semantic validation, anomaly detection, and powering a natural language query Slack bot.

#9about 1 minute

Why data governance is the key to unlocking value

The key takeaway is that more data often leads to more confusion, and implementing strong data governance is the only way to create clarity and value.