I’ve been speaking with leaders in the energy and utilities market about data transformation and the accelerated journey to AI adoption.
Something which stood out to me is that the conversation has moved on from how businesses can use AI to something much more practical. Decision-makers are asking if their data and operating model are AI-ready.
From those conversations, five challenges keep coming up.
What are the top 5 challenges the energy and utilities market is facing?
- Data is still fragmented across the organisation
- AI is exposing weaknesses in data quality
- Turning data into operational insight
- Scaling AI beyond isolated use cases
- Finding people who understand both the technology and the industry
Data is still fragmented across the organisation
The problem is more nuanced than simply having ‘legacy systems’.
Asset, customer, operational, GIS, finance and trading data often sits across different platforms, teams and ownership structures. The challenge is getting that information to work together consistently enough to support enterprise-level decision-making. Several of the household name companies are undergoing large transformation projects to unify messy data sources into a single point of truth, with governance surrounding it to determine who can access which parts of the warehouse.
This is becoming increasingly important as the energy system becomes more decentralised and interconnected. Ofgem’s current work on data domains, interoperability and the Data Sharing Infrastructure reflects exactly this issue. Transparency has never been more important. Not to mention the scrutiny from the public as we continue through a time of increased energy bills and climate goals.
AI is exposing weaknesses in data quality
There is a temptation to treat AI as a technology problem.
Many organisations are discovering that the harder question is whether the underlying data is accurate, accessible, well-governed and contextualised.
Poor master data, inconsistent definitions, missing metadata and unclear ownership don't disappear when you introduce an LLM or machine-learning model. They become more visible.
I have noticed system integration, data quality and data access among the most common technical challenges.
Turning data into operational insight
Energy companies have enormous amounts of data. The challenge is increasingly making it useful at the point a decision needs to be made.
That could mean understanding network constraints, forecasting demand, managing distributed assets, improving asset maintenance or responding to changing customer behaviour.
The UK's energy system is becoming more weather-dependent and operationally complex. This increases the requirement for better forecasting, optimisation and flexibility.
The gap isn't necessarily a lack of data. It's the engineering, architecture and domain expertise required to turn that data into something operational teams can actually use. We are seeing smaller energy businesses move from manual reporting into much clearer visualisation - using tools like PowerBI within Fabric architecture.
Scaling AI beyond isolated use cases
There are plenty of successful AI experiments. The harder part is moving from a promising proof of concept to something that can operate reliably within a regulated, safety-critical environment.
That requires much more than an AI model. Organisations need the right data pipelines, cloud architecture, integration, monitoring, governance and people around it.
Ofgem's recent work on AI assurance specifically highlights inconsistent governance, limited real-time monitoring and difficulties integrating AI into existing operational and safety frameworks.
Finding people who understand both the technology and the industry
This is probably the challenge I hear most often.
Energy and utilities don't only need more people who are experienced in Python, Azure, Databricks or LLMs. They need people who can understand why the data matters to the business and translate between engineering teams, operational stakeholders and senior decision-makers.
We are seeing huge value in professionals who bring both the technical skill set to productionise AI models and platforms with the power skills to influence, communicate, sell and present to the board.
As data transformation becomes increasingly embedded into the core operating model, the value of people who can bridge those worlds becomes significant.
The bigger picture
I don't think the next phase of energy and utilities transformation is primarily about finding the next AI tool.
It is about building the data foundations, architecture and capability that allow these technologies to deliver value safely and at scale.
The direction of travel from Ofgem and government is increasingly clear: better data quality, accessibility and governance is fundamental to a more digitalised energy system.
The organisations that make the most progress won't necessarily be those that adopt the most AI solutions and services, it will be the ones that can combine
strong data foundations + the right technology + people who understand how to apply both to the realities of the energy system.
That is where I’m seeing some of the most interesting conversations happening right now.
If you are discussing your data journey and identifying how you get from A to B, reach out. I’d be thrilled to set up a chat with our internal experts and share some stories from the market.
