Changing the game: what happens when anyone can talk to their energy data? 

Energy data used to belong to specialists. But with AI moving at record speed, things are changing. We sat down with EG Software Product Manager Iiro Kuntsi to discuss what happens when controllers, maintenance staff and new hires can ask their energy data questions – and why products should be built together with its intended users. 

Energy data has always belonged to the few people trained to read it. The AI in EG EnerKey opens it to everyone else in the organization – the controller who needs the figures for a report, the facility manager who wants to know why a building spiked last night, or the CFO analyzing consumption without waiting for the energy specialist. The first part of this series was about the condition that makes this acceptable: the data stays exactly as protected as it was. This part is about the other condition. New functionality should be developed in collaboration with its users, if we want software that solves real problems.  

For the last two decades, energy management systems have been a specialist's tool. Everyone else asked the specialist or waited for the report. That is the habit the AI in EG EnerKey was built to break. Iiro comments:  

“Customers consistently describe our new AI-agent as intuitive. It removes the learning curve of a traditional energy management system. Why spend time learning a complex system, when you can just ask it a question? For many of them it's simply about making EG EnerKey more accessible for new staff and users who aren't seasoned energy experts.”  

What failed AI pilots teaches us – and why AI must be built together with its users 

Most enterprise AI fails. MIT researchers who studied hundreds of generative-AI pilots in 2025 found that 95 percent produced no measurable business effect, and concluded that the barrier was not technology or talent, but learning – the systems did not adapt to the people using them. 

This insight is essential. Without close collaboration with the future users, software has a high risk of failing. Iiro continues: 

“Nobody had built a similar AI-agent for energy management before – there was no playbook to follow. We didn't want to develop it in isolation and hand our customers a finished product. We launched early, with selected restrictions on data and load, specifically so we could learn from real usage with real users, before scaling further.”  

By August, more than 50 customer organizations had activated the AI – retail, telecom, facility management, real estate – and thousands of questions had been asked.  

Energy specialists asked it to scan whole portfolios for elevated night-time base load or heating behaving abnormally. Controllers wanted monthly consumption and emissions reports that produce themselves. Electricity balance analysts asked whether it could read tariff documents from the EG EnerKey archive and compare them with consumption to find fuse-size and tariff savings across hundreds of sites.  

McKinsey's 2025 survey found that the strongest predictor of value is redesigning the workflow around the AI, rather than bolting AI onto the old process. Iiro again: 

“The question is never ‘what can the AI generically do’. Our customers are asking it to simplify or automate a specific job for a specific role. We’re packaging those role-based use cases as named workflows – for the analyst, the specialist, the maintenance manager, the portfolio director, the CEO – rather than leaving it as one generic agent that everyone has to figure out how to use for their own job.” 

Keeping customers inside the development process 

Real use also showed the limits. Some early questions on very large portfolios asked more than the AI could yet deliver, and the team made performance at scale its priority through the end of the year. Releases ship every two weeks, so what it can do changes monthly. 

“Communicating what's live today versus what's coming next is part of how we keep customers inside the development process with us, rather than handing them a black box and asking them to wait. Broad enthusiasm is there – but adoption has to be earned, not assumed, by every role in our customers' energy management process.” 

The next development stage for the AI-agent is generation. Charts in the interface and exportable reports built from live data. During 2027, the plan is for the AI to propose changes – to facilities, meters, alarms and actions – with the user approving each one. The destination came from a customer before it was on the roadmap. Iiro specifies: 

“The AI notices a deviation, comments on the probable root cause, and the finding goes automatically to the facility management system through EG PropCore. Action is taken on site, and the status flows back into EG EnerKey. A customer described exactly this to us. It's a nice confirmation that we're building toward the same outcome they're imagining.” 

The future of energy management is built together 

Iiro concludes his approach to AI development:  

“Our philosophy is to be at the AI frontline together with our customers, simplifying and automating the parts of energy management that are still manual, or inaccessible. That means making the AI more capable and the experience better incrementally – every month, every release.” 

AI on its own is just a tool. But when it allows anyone to talk to their building’s energy data, in a safe environment, magic happens.  

Already an EG EnerKey customer? The AI can be added to your existing plan – contact your account manager.

Related products