AI success starts with clean data, not just better models | Databricks Blog
Skip to main content
Login
Why Databricks
Discover
For App Developers
For Executives
For Startups
Lakehouse Architecture
Databricks AI Research
Customers
Customer Stories
Partners
Partner Overview
Explore the Databricks partner ecosystem
Partner Spotlight
Featured partner announcements
Partner Program
Explore benefits, tiers and how to become a partner
Cloud Providers
Databricks on AWS, Azure and GCP
Find a Partner
Discover Databricks partners for your needs
Partner Solutions
Find custom industry and migration solutions
Product
Databricks Platform
Platform Overview
A unified platform for data, analytics and AI
Data Management
Data reliability, security and performance
Sharing
Open, secure, zero-copy sharing for all data
Data Warehousing
Serverless data warehouse for SQL analytics
Governance
Unified governance for all data, analytics and AI assets
Data Engineering
ETL and orchestration for batch and streaming data
Artificial Intelligence
Build and deploy ML and GenAI applications
Business Productivity
Unified search, chat, dashboards and apps
Business Intelligence
Intelligent analytics for real-world data
Application Development
Quickly build secure data and AI apps
Database
Postgres for data apps and AI agents
Security
Open agentic SIEM built for the AI era
Integrations and Data
Marketplace
Open marketplace for data, analytics and AI
IDE Integrations
Build on the Lakehouse in your favorite IDE
Partner Connect
Discover and integrate with the Databricks ecosystem
Pricing
Databricks Pricing
Explore product pricing, DBUs and more
Cost Calculator
Estimate your compute costs on any cloud
Open Source
Open Source Technologies
Learn more about the innovations behind the platform
Solutions
Databricks for Industries
Communications
Media and Entertainment
Financial Services
Public Sector
Healthcare & Life Sciences
Retail
Manufacturing
See All Industries
Cross Industry Solutions
AI Agents
AI Governance
Cybersecurity
Marketing
Migration & Deployment
Data Migration
Professional Services
Solution Accelerators
Explore Accelerators
Move faster toward outcomes that matter
Resources
Learning
Training
Discover curriculum tailored to your needs
Databricks Academy
Sign in to the Databricks learning platform
Certification
Gain recognition and differentiation
Free Edition
Learn professional Data and AI tools for free
University Alliance
Want to teach Databricks? See how.
Events
Data + AI Summit
Data + AI World Tour
AI Days
Event Calendar
Blog and Podcasts
Databricks Blog
Explore news, product announcements, and more
AI Blog
Explore our AI research and engineering work
Data Brew Podcast
Let’s talk data!
Champions of Data + AI Podcast
Insights from data leaders powering innovation
Get Help
Customer Support
Documentation
Community
Dive Deep
Resource Center
Demo Center
Architecture Center
About
Company
Who We Are
Our Team
Databricks Ventures
Contact Us
Careers
Working at Databricks
Open Jobs
Press
Awards and Recognition
Newsroom
Security and Trust
Security and Trust
DATA + AI SUMMIT
JUNE 15–18|SAN FRANCISCO
Join us at the world’s largest data, apps and AI event.
Register
Ready to get started?
Get a Demo
DATA + AI SUMMIT
JUNE 15–18|SAN FRANCISCO
Join us at the world’s largest data, apps and AI event.
Register
Login
Try Databricks
All blogs
/
Industries
Table of contents
Why data unification is table stakes
The number no one trusts
AI is the forcing function enterprise analytics needed
Metadata as the missing ingredient
From monthly reports to hourly decisions
Getting people into the data
The advice most C-suite get wrong
What good looks like from here
Closing Thoughts
Table of contents
Table of contents
Why data unification is table stakes
The number no one trusts
AI is the forcing function enterprise analytics needed
Metadata as the missing ingredient
From monthly reports to hourly decisions
Getting people into the data
The advice most C-suite get wrong
What good looks like from here
Closing Thoughts
Energy
May 5, 2026
AI success starts with clean data, not just better models
Why building an AI-ready foundation doesn't stop at the technology
by Aly McGue
Summary
Platforms like Kraken and Databricks solve the foundational data challenge, giving organizations unified, well-documented data that makes everything from self-service analytics to AI viable.
Once you solve the unified data challenge, the rest is a business problem.
Data is a business asset, not an IT platform. The organizations pulling ahead pair unified data with deep business context and a data-literate culture.
Kraken, the AI-powered operating system behind some of the world's largest utilities, manages over 90 million customer accounts across 27 countries for clients including EDF, E.ON, National Grid, and Tokyo Gas. Kraken uses Databricks as its internal data platform and partners with Databricks to help clients maximize the value of the data they receive via secure, scalable data distribution.
Kristy Mayer-Mejia is the Global Head of Data Transformation at Kraken, where her team helps utility clients understand, adopt, and extract value from the data Kraken provides. Her mandate is twofold: speed up the time it takes clients to use the data, and increase the value they get from it.
I sat down with Kristy to understand how data functions as a business asset and is the foundation of a successful AI strategy. A key point from our conversation is that becoming data-driven is as much about clean, unified data as it is about deep business context and ownership. Platforms like Kraken and Databricks solve what Kristy calls the foundational unification problem, the prerequisite that makes everything else viable. But once that foundation is in place, the part most leaders underestimate is the business context that makes unified data usable.
Why data unification is table stakes
Aly McGue: In your experience, why do siloed data and fragmented systems remain a big hurdle for organizations trying to extract value from their investments?
Kristy Mayer-Mejia: What we see repeatedly with our clients is that low-quality, siloed data is the single biggest blocker to getting value from any other investment. Until the data is in one place, nothing else works at scale, and solving that is exactly the problem Kraken’s platform is designed to address. And I've lived this as a data leader in all my prior roles, too. Your team spends 80% of their time cleaning data, and that's just not valuable work. It's not necessary.
The real unlock is self-service, but it’s only possible once the underlying data is clean, unified, and accessible. Especially in the age of AI, self-service is possible at scale. You're never going to move quickly as a business, innovate or make day-to-day data-driven decisions if every question has to be answered by the data team. But when the data is scattered across systems with no documentation and no clear way to join it, self-service is impossible. Unification is this foundational unlock that makes everything else viable: the analytics, the AI, the speed of decision-making. It's table stakes.
The number no one trusts
Aly: We’ve all been in meetings where leadership spends more time debating 'which number is right’ than actually making a decision. What is the hidden cost of that lack of trust in the data?
Kristy: I give this example all the time, and it's been true at every company I've ever worked at. Before you have unified data, the classic question is: how many customers do we have? And no one totally knows. You know the rough magnitude. But when I give that example, every time people laugh because they know it's true.
What it leads to is a lack of trust in the data. And one of the primary early values that unified data provides is the speed of decision-making, the ability to embed data-driven thinking in the company's DNA. You can't move quickly if every time you pull a number, you're thinking, ’ Am I sure this is right? ’ Let me check five other places. Let me ask someone. And then it's different. And then you have to run down why it's different. Suddenly, it's two weeks later or a month later, and you might as well have just picked a random direction and kept moving.
AI is the forcing function enterprise analytics needed
Aly: We often talk about data fueling AI, but you’ve suggested that AI might actually be a 'forcing function’ for better data. How is the push for AI changing the way organizations approach documentation and context?
Kristy: AI has actually been a forcing function. The inputs AI needs are the same inputs humans need: clear data, documentation, context on what columns mean, and how things join together. When data is hard to use, self-service analytics feels like a nice-to-have because the value is hard to pin down. It's a few hours saved here and there on individual decisions, which doesn't feel compelling in isolation. But accumulated across the organization, it's huge. It's just hard to see.
AI has made that value visible and has made clean data & documentation table stakes. It takes what everyone always knew was needed and makes it non-negotiable. And then on the other side, AI itself provides the tools to unlock analytics. Things like conversational interfaces that let people query data without writing SQL. So it's both the forcing function that drives unification and the payoff that comes out of it.
Metadata as the missing ingredient
Aly: You've talked about the need to unify and document data. But when it comes to AI specifically, is documentation in a knowledge base or a PDF enough?
Kristy: It used to be. We shared our data documentation the way most companies do: a PDF, or a page on a website that a data analyst could reference when they needed context. That works well enough for humans. It does not work for AI.
Every client I talk to now is asking the same question: can you share the metadata in context, alongside the data itself, so we can actually feed it into models and have them understand what they're working with? That shift, from documentation as a reference artifact to documentation as a live input, is one of the more underappreciated changes AI is forcing. With Unity Catalog and Delta Sharing, we can share that context with the data rather than separately from it. For our clients, that is often the difference between AI that can reason about the data and AI that cannot.
From monthly reports to hourly decisions
Aly: What does 'data unification’ look like in practice? How does near-real-time visibility change day-to-day operations?
Kristy: A few examples from our clients stand out. One is call center operations, which is a massive function for utilities. We had a client go from monthly reporting on call volume, which was so painful to put together, to dashboards that update every couple of hours, with a predictive model layered on top of what calls they're likely to see going forward. That ability to fine-tune operations in near real time, rather than looking backward once a month, is a completely different way of running the business.
Another area is product innovation. In the utility space, clients are determining which products and tariffs to offer to attract and retain customers. That's a decision that can be deeply optimized with data. Clean, clear data give clients easy insight, and rapid test-and-learn cycles to optimize their product offers – and then Kraken's platform lets them quickly launch those new tariffs.
Getting people into the data
Aly: The 'analyst bottleneck’ is a classic pain point for leadership. How do natural language interfaces, like Databricks Genie, shift the culture from waiting weeks for a report to getting answers in minutes?
Kristy: Most of our Genie clients are still in the early stages. But what we're seeing is that it's accelerating their time to get started by weeks or more. They don't need to deeply model the data the way you would to feed it into a traditional BI tool. They need clear documentation, they need the context, they need the data in one place, but they don't have to structure it so precisely that a user can explore it through a rigid interface.
But beyond the speed, there's a really clear cultural knock-on effect. One of the bigger barriers to data value is the cultural shift of making data part of your DNA. And I firmly believe one of the keys to that is making it incredibly easy and intuitive. When the barrier is low, and people can get in quickly, the culture, and the compounding value, follows.
The advice most C-suite get wrong
Aly: What is the biggest misconception C-level leaders have when they task their IT departments with 'getting the data ready’ for AI?
Kristy: Data is a business asset. And the biggest mistake I see leaders make is treating it like an IT platform. They disconnect it from the business and say, "Okay, IT, go prepare our data.” But the key to building a solid data foundation is the deep business context. How is the data generated? How is it used? How do people interpret it? What does this field actually mean? Once the technical foundation is in place, the hardest part becomes that deep business context. And the vast majority of that work sits with the business, not the data team.
So my advice is to embed data within the business. The roadmap to getting your data ready for AI is a shared roadmap. It's a business roadmap as much as it is a technical one.
What good looks like from here
Aly: Kraken sits across a large share of the utility industry's data. Where do you see AI and data taking your clients over the next three to five years?
Kristy: What I find most interesting is how quickly AI is raising the ceiling on what clients can do once they have a solid data foundation. For a long time, the question was: how do we get our data into a usable state? That work is still real, and it still takes time. But the question is shifting toward: now that the foundation is there, what becomes possible? And the answer to that keeps expanding. AI is changing where clients start from and what good looks like. Clients who would have considered a monthly report a success two years ago are now running hourly dashboards with predictive models layered on top and looking quickly toward broad use of agentic AI.
The ones who invested early in their data capabilities – and not just their tech but their skills and culture – are the ones moving fastest now, and the gap between them and everyone else is only going to widen.
Closing Thoughts
Kristy's perspective adds an often-missing layer to the data infrastructure conversation. The platform and the unification it enables are the foundational unlock. But where she sees most organizations stall is in the work that comes after: the business knowledge that makes data usable, the documentation that makes AI possible, and the cultural shift that makes self-service real.
As you develop your roadmap to embed AI across your organization and products, download the Databricks State of AI Agents to help you benchmark your investments.
Get the latest posts in your inbox
Subscribe to our blog and get the latest posts delivered to your inbox.
Sign up
View all blogs
Why Databricks
Discover
For App Developers
For Executives
For Startups
Lakehouse Architecture
Databricks AI Research
Customers
Customer Stories
Partners
Partner Overview
Partner Program
Find a Partner
Partner Spotlight
Cloud Providers
Partner Solutions
Why Databricks
Discover
For App Developers
For Executives
For Startups
Lakehouse Architecture
Databricks AI Research
Customers
Customer Stories
Partners
Partner Overview
Partner Program
Find a Partner
Partner Spotlight
Cloud Providers
Partner Solutions
Product
Databricks Platform
Platform Overview
Sharing
Governance
Artificial Intelligence
Business Intelligence
Database
Data Management
Data Warehousing
Data Engineering
Business Productivity
Application Development
Security
Pricing
Pricing Overview
Pricing Calculator
Open Source
Integrations and Data
Marketplace
IDE Integrations
Partner Connect
Product
Databricks Platform
Platform Overview
Sharing
Governance
Artificial Intelligence
Business Intelligence
Database
Data Management
Data Warehousing
Data Engineering
Business Productivity
Application Development
Security
Pricing
Pricing Overview
Pricing Calculator
Open Source
Integrations and Data
Marketplace
IDE Integrations
Partner Connect
Solutions
Databricks For Industries
Communications
Financial Services
Healthcare and Life Sciences
Manufacturing
Media and Entertainment
Public Sector
Retail
View All
Cross Industry Solutions
AI Agents
AI Governance
Cybersecurity
Marketing
Data Migration
Professional Services
Solution Accelerators
Solutions
Databricks For Industries
Communications
Financial Services
Healthcare and Life Sciences
Manufacturing
Media and Entertainment
Public Sector
Retail
View All
Cross Industry Solutions
AI Agents
AI Governance
Cybersecurity
Marketing
Data Migration
Professional Services
Solution Accelerators
Resources
Documentation
Customer Support
Community
Learning
Training
Certification
Free Edition
University Alliance
Databricks Academy Login
Events
Data + AI Summit
Data + AI World Tour
AI Days
Event Calendar
Blog and Podcasts
Databricks Blog
AI Blog
Data Brew Podcast
Champions of Data & AI Podcast
Resources
Documentation
Customer Support
Community
Learning
Training
Certification
Free Edition
University Alliance
Databricks Academy Login
Events
Data + AI Summit
Data + AI World Tour
AI Days
Event Calendar
Blog and Podcasts
Databricks Blog
AI Blog
Data Brew Podcast
Champions of Data & AI Podcast
About
Company
Who We Are
Our Team
Databricks Ventures
Contact Us
Careers
Open Jobs
Working at Databricks
Press
Awards and Recognition
Newsroom
Security and Trust
About
Company
Who We Are
Our Team
Databricks Ventures
Contact Us
Careers
Open Jobs
Working at Databricks
Press
Awards and Recognition
Newsroom
Security and Trust
Databricks Inc.
160 Spear Street, 15th Floor
San Francisco, CA 94105
1-866-330-0121
See Careers at Databricks
© Databricks 2026. All rights reserved. Apache, Apache Spark, Spark, the Spark Logo, Apache Iceberg, Iceberg, and the Apache Iceberg logo are trademarks of the Apache Software Foundation.
Privacy Notice
|Terms of Use
|Modern Slavery Statement
|California Privacy
|Your Privacy Choices