AI is rapidly evolving from being a novel concept to a business necessity. In our recent UK and Ireland event, leaders from the Energy and Utilities sector gathered to explore how AI can drive immediate, transformative change.

AI has moved beyond the hype cycle and proven its value. The application of generative AI in particular, holds the potential to significantly boost both the bottom line and the top line for businesses. The multimodal capabilities of large language models (LLMs), which are the brain behind generative AI, make human language the new programming language. This technology can unlock the value in dormant data, enhance productivity by integrating fragmented application landscapes, and open up new horizons for driving effectiveness and innovation.

I was also joined by my fellow hosts, Mark Corley, Energy and Utilities Global Lead at Avanade, and Microsoft AI Director, Frederic Goursaud. The event gave us the opportunity to engage with our clients – CxOs, heads of trading desks, and veteran data and technology leaders – and discuss the business value drivers for applying generative AI to tackle the industry's challenges. We explored how this technology can truly redefine the future, and the discussions were highly insightful.

Transition to net zero: Decarbonizing the energy grid
The Energy and Utilities market faces its own unique set of circumstances. The UK has set an ambitious target of achieving a clean electricity grid by 2030, according to government plans. This goal necessitates the decommissioning of carbon-heavy assets, which will be partially replaced by renewable energy sources like wind and solar. However, these renewables are intermittent, posing significant challenges for power generation and planning.

It's astonishing that, due to grid constraints, we sometimes have to turn down wind generation and instead pay Combined Cycle Gas Turbines (CCGTs) to come online. This situation highlights the increasing importance of battery storage. Batteries not only stabilize the grid by storing excess energy from renewables, but also generate revenue through tolling agreements and ancillary services.

The ability to extract actionable insights from data is crucial. Without it, predicting the impact of weather, pricing and market forces, will become increasingly difficult for Energy and Utilities providers and the businesses they serve. Advanced data analytics and AI-driven solutions are essential to navigate these complexities and ensure a reliable and efficient energy supply.

At the same time, end customers are making more sustainable choices, such as switching to heat pumps and becoming energy providers themselves. The market is undergoing a transformation from top to bottom, requiring the industry to adapt with robust digital architectures that can support this shift toward sustainability and resilience.

Generative AI: Revolutionizing data visibility and insights
These foundational market changes challenge the forms and functions of traditional energy trading. You need to be able to tap into data insights in real time, to see how all these factors bear out second by second. Regardless of how talented traders may be, any headway is impossible to achieve without tools that can analyze data at scale. And now disrupters are leveraging real-time data insights to scale algorithmic trading. These automated trading processes and LLMs are absolutely fundamental to understanding and acting upon the information coming from the fast-paced energy markets.

Taking a pragmatic approach to implementation
As I tabled on the day, Avanade stands at the forefront of generative AI innovation in the Energy and Utilities sector. Our approach emphasizes business engagement, quantifying tangible value, and addressing transformation with a modular strategy. This means we prioritize proving concepts first and scaling them later.

One of the most significant risks we have identified for organizations is the potential misuse of ChatGPT with enterprise intellectual property (IP). To mitigate this risk, it is crucial for organizations to implement generative AI responsibly within their deployments. This includes establishing robust guardrails for enterprise security within their cloud platforms, ensuring that data integrity and confidentiality are maintained.

When you have a secured foundation for AI within your enterprise, you can infuse your own data into the LLMs. This approach is called RAG (Retrieve Augment Generate).

RAG improves LLMs by incorporating additional data. While LLMs can reason broadly, their knowledge is limited by the publicly available data at the time they’re trained – so they can’t adapt in real time. RAG allows AI applications to access and use newer data, by retrieving relevant information from private sources and incorporating it into model prompts, to reflect the real-time market state. By using RAG, we can prevent AI hallucinations, maintain control over knowledge sources, lower the computational and financial costs of running LLM-powered chatbots, and expand use cases.

RAG categorizes generative AI applications into four main use case groups, namely:
  • Chat-based solutions
  • Business process automation
  • Multi-agent systems
  • Model fine-tuning

This framework, alongside the scale and experience that Avanade can provide, enables us to embed the systematic change that the Energy and Utilities industry needs to respond to the influx of new data and market influences. As we look to the future, RAG is a valuable foundation for the next wave of generative AI innovation.

We’ll see networks of generative AI agents, simultaneously performing data collection, curation and execution, unconstrained by manual intervention. And as the technology evolves, we’ll be able to build multi-agent frameworks, reproducing trading organizations with virtual agents across compliance, governance, finance and beyond.

Microsoft Fabric: Unifying data and unlocking value
Frederic Goursaud, Director of Microsoft's specialist AI unit, shared Microsoft's AI roadmap, as well as the future direction of travel in AI platform innovation. This included exciting new features of real-time analytics in Microsoft fabric.

Microsoft Fabric is a turnkey SaaS platform for user-specific analytics. Users can quickly set up a notebook environment, access data and create a working setup that fits their needs. They can schedule automatic data refreshes with a few clicks, and easily share or mark data as trusted for colleagues to use. In Energy and Utilities, the platform supports advanced analytics and data structures. As an example, it can bring together machine learning, regression models, weather forecasting and tools like GitHub Copilot to execute data analytics and simulations that can inform decisions. In addition, Microsoft Fabric is built on open standards, so it is compatible with a huge range of other frameworks. It integrates with the existing data environment as a new, intelligent management layer, unlocking more value without having to rip and replace infrastructure.

As a trusted provider to the Energy and Utilities industry, we can deliver a coherent vision for generative AI. And we can prove it as well. In our live demonstration on the day, we were able to show how we could gather data from Elexon, the National Grid, and third-party providers with Microsoft Fabric, in just 30 minutes. It was created using analytics to build a real-time weather forecasting model, with simulations used to train the model and back-test results. These curated datasets could be used to underpin algorithmic trading and enable self-service reporting with Microsoft Copilot.

But how do we answer the question of implementation? A common experience for businesses looking to trial algorithmic trading is that they continually try to hone and specify datasets. What started as a noble goal becomes an endless series of workshops, a situation that we call ‘analysis to paralysis’. However, with a full-stack solution, you can test and prototype immediately, to avoid uncertainty and apply your data with a provable benefit.

We can even see the value of Microsoft Fabric in our own experiences. It is the technical backbone of Avanade’s own data strategy, used to improve data modelling and analytics. Microsoft Fabric provides internal reporting and insights that are used by over 60,000 people, enabling us to optimize our approach to data and support the work our employees do every single day.

Working with a partner that gets you where you want to go
Avanade is pioneering all aspects of generative AI across the Energy and Utilities value chain. While many consultancies are targeting proofs of concept (PoC) to show the value of AI, we believe that PoC won't deliver the fundamental commercial reinvention that generative AI promises. Instead, we have assembled a comprehensive framework to assess impact, prioritize AI projects, and successfully realize returns that measurably improve productivity and growth for Energy and Utilities clients.

In addition, Avanade’s work is recognized throughout the industry. We have recently been named as Microsoft's first ever Global Business Transformation Partner of the Year for Copilot alongside Accenture, as well as the Best AI Consulting Service Provider in the AI Breakthrough Awards 2024. We are also Partner Zero for Microsoft Fabric. From the very start, we’ve worked to develop and deploy the technologies that will help industries to unlock value from data and create the competitive advantages they need to stay ahead.

If you want to explore ways that AI can enhance your business, but you’re not sure of the best approach to test, learn and scale the technology, Avanade can help.

In our AI Transformation Studio, we provide the expertise and frameworks to help you confidently gain velocity on AI, driving value early and giving you a platform to accelerate innovation.

Find out more about our work with the Energy and Utilities industry using the link below.