Utility Warehouse: Boosting customer and partner experiences with a data mesh, analytics, and AI

About Utility Warehouse

With over 800,000 customers, Utility Warehouse is one of the UK's leading multiservice utilities providers, helping customers save time and money by combining their energy, mobile, broadband, and insurance bills into one account, thus offering savings with the more services they take.

Industries: Energy, Chemicals & Utilities
Location: United Kingdom

Tell us your challenge. We're here to help.

Contact us

Utility Warehouse used Google Cloud and its own tool Heimdall to build a cloud data platform that forms the technological underpinning of a wider cultural shift to data, analytics, and AI-based decision-making.

Google Cloud results

  • Provides the foundation for a data-centric, responsive company culture with BigQuery and Looker serving 200+ users
  • Develops deeper insights into customer and partner needs with AI tools supported by Dialogflow and Vertex AI
  • Reduces the time to production for a quicker, more effective response to changes in the market using efficient reporting and analytics
  • Provides the capability of deploying and monitoring model at scale with Google Cloud and in-house tool Heimdall

Slashes time to production with BigQuery and Vertex AI

Time is precious. Spare time, even more so. It should be spent with family, friends, or simply enjoying oneself. But too often, it's spent fretting about energy or internet bills or wading through endless price comparison sites to find the best deals. As the cost of living has risen dramatically due to unstable energy markets, managing utilities has become a major source of concern.

As a multiservice provider, Utility Warehouse helps customers save time and money by bundling together their energy, mobile, broadband, and insurance packages into one account with one monthly bill. And by doing this, customers make savings on their monthly bills.

The company also helps customers to earn commission by signing up as "Partners" and introducing Utility Warehouse to their own community. By working with more than 50,000 Partners, the company is now serving more than 800,000 customers.

As the company continues to expand, every part of the business is focused on making sure that existing customers have the best possible experience.

"We wanted to accelerate the culture of analytics throughout the company and become much more data-driven in how we make our decisions and go about things. Building a data platform with Google Cloud was pivotal to this process."

Didier Vila, Chief Data and AI Officer, Utility Warehouse

"We want to maximize our customer growth, and the way we do that is by maximizing our customer and partner experience," explains Didier Vila, Chief Data and AI Officer, Utility Warehouse. "Within data, tech, product and commercials teams, we ask ourselves daily how to translate and solve these questions."

Utility Warehouse has an ambitious target of adding 1 million new customers over the next four to five years. As it strives to improve its service and reach new customers, the company has accelerated its data journey, and begun to explore the possibilities opened up by cutting-edge analytics and artificial intelligence. The technological underpinning for all of this has been built with Google Cloud.

"We wanted to accelerate the culture of analytics throughout the company and become much more data-driven in how we make our decisions and go about things," says Didier. "Building a data platform with Google Cloud was pivotal to this process."

Constructing a data mesh

Since its founding in 2002, Utility Warehouse has continually evolved its stack, resulting in a multicloud infrastructure that enables scale and flexibility. But there's a possibility of creating data silos, which limits collaboration across the company and operational effectiveness.

A crucial part of the company's quest for a data-led culture is the concept of the "data mesh." At Utility Warehouse, the mission of the data infrastructure team is to develop framework and data product archetypes that enable all departments to create their own "data products" and expose them to other parts of the business, resulting in a distributed but all-encompassing data architecture that is faster and more flexible than traditional centralized repositories.

"BigQuery is the heart of our data solution. It is where we collect all the data into one place, forming the central component of our ecosystem."

Didier Vila, Chief Data and AI Officer, Utility Warehouse

Moreover, Utility Warehouse brings all of its analytics and operational datasets together on BigQuery, with pipelines built with Dataform for efficient and scalable "Extract, Transform, and Load" operations.

"BigQuery is the heart of our data solution" says Didier. "It is where we collect all the data into one place, forming the central component of our ecosystem."

With all of its data collated onto a single platform, Utility Warehouse can then begin to mine it for insights. A natural step has been the democratization of data, with the company using Looker to enable colleagues from every department to analyze data held in BigQuery and rapidly create their own reports and dashboards for enhanced decision-making. Utility Warehouse has also developed and applied artificial intelligence to its data, using elements of Dialogflow and Vertex AI in support of its own proprietary tools and models, in particular, Heimdall.

Building Heimdall

The north star for the data team is to scale AI at Utility Warehouse. The company aims to build more products and decision systems with ML/AI at the core. As time passes, things change, including data and machine learning models. As Utility Warehouse continues to scale its AI capabilities, it will need a way to monitor its machine-learning models and ensure that they operate as effectively as possible on high quality data. To do this, Utility Warehouse has built a custom model monitoring tool called Heimdall.

Heimdall is built with the aim of being scalable, customizable, and innovative, providing a one-stop-shop for ML/AI model monitoring. Heimdall serves both the data and ML practitioners with its deep insights into models running in production. At the same time it also helps those in product and business facing verticals by providing an overview of key metrics with its dashboard style UI.

Heimdall runs on Google Kubernetes Engine and seamlessly integrates with Vertex AI and other open source tools, quickly becoming a vital part of the Utility Warehouse MLOps stack.

Embedding data into core business decisions

The new platform and AI tools are all part of a much wider effort to democratize data throughout the company. The uptake of Looker as a Business Intelligence reporting tool facilitated this adoption. Today, more than 200 users throughout the company in every department use Looker to generate daily, actionable insight about its customers, and partners, and the wider market environment.

"Building a data mesh with Google Cloud means that we can share data and insights across the company seamlessly," says Didier. "We're more connected, and we can work together to find holistic solutions."

"Building a data mesh with Google Cloud means that we can share data and insights across the company seamlessly. We're more connected, and we can work together toward holistic solutions."

Didier Vila, Chief Data and AI Officer, Utility Warehouse

This access to data has changed the rhythm of the way Utility Warehouse colleagues work. "Google Cloud has helped us to significantly reduce our production lifecycle," says Didier. "We can take decisions and respond much more quickly and effectively to sudden changes, such as the recent introduction of an energy price cap by the government or the market stabilization charge scheme."

The new data platform has also paved the way for a flourishing list of AI initiatives among these two. The first is the capability to discover and learn the behavior and needs of its partners. "It helps us understand the needs of our 50,000 partners and improves our communication with them," explains Didier.

The second initiative is a natural language processing model, which looks for emerging topics in customer communications and feeds that back to the customer experience team. This means that they can very quickly understand and respond to new trends in the utility markets and do their best to adapt the service to customer needs.

This is an exciting collaborative journey for Didier and his team at Utility Warehouse. Alongside building a technological foundation for its data platform, the company has also been working on improving our processes, upskilling colleagues, and learning together towards data. In particular, the company partnered with Cambridge Spark, to accelerate this learning journey. "We are becoming more data-centric as a company every day," says Didier. "There is a huge appetite for data-driven decisions, and AI use cases across the company. We're excited to work with all our talented colleagues to accelerate our data-AI journey."

Tell us your challenge. We're here to help.

Contact us

About Utility Warehouse

With over 800,000 customers, Utility Warehouse is one of the UK's leading multiservice utilities providers, helping customers save time and money by combining their energy, mobile, broadband, and insurance bills into one account, thus offering savings with the more services they take.

Industries: Energy, Chemicals & Utilities
Location: United Kingdom