AI in clean energy: This company wants you to be able to talk with your projects

AI in clean energy: This company wants you to be able to talk with your projects

To make sense of what's what in AI, Albert Hofeldt, the senior vice president of technology for solar and battery storage software provider Stem, joined Episode 65 of the Factor This! podcast. Subscribe wherever you get your podcasts.

AI is everywhere. It's impossible to avoid. But parsing revolutionary technology from marketing hype has never been more challenging.

So, what does AI really mean for clean energy? Beneath the frothy layer of pretenders and contenders lie potentially game-changing applications for grid management, asset monitoring and optimization, and market participation.

One company, Stem, believes asset owners should soon be able to communicate with their projects. The solar and storage software company is building out a language model that, paired with a digital twin, could allow for two-way communication between clean energy pros and their infrastructure in the field.

To make sense of what's what in clean energy AI, Stem's senior vice president of technology, Albert Hofeldt, joined the Factor This! podcast to break down the company's plans and how to separate winners from losers in AI.

"There's quite a bit of hype in this whole cycle," Hofeldt said. "What sometimes are basic correlations and relatively simple analytics, you're finding that people are just blowing this stuff out of the water and calling it AI."

Hofeldt has spent a career analyzing and building AI and machine learning tools, first with Deloitte, then Thomas Reuters, Wood Mackenzie, financial technology firm LiquidX, and, now, Stem.

The clearest way to assess an AI tool's viability, Hofeldt said, is to first examine its backbone: data. While some innovative solutions are bound to begin at zero, often data history will tell the story.

"Look across the past six months, 18 months, 24 months, and look for some level of connective tissue in the statements they're making," he said. "Quite often you'll see organizations popping up with these amazing offerings, but there's no history there."

Watch the full episode on YouTube

When Stem first launched in 2009, the company focused on providing customers with a turnkey storage solution that included battery hardware and, later, software to optimize the asset. It was a "black box," by design.

But developers and asset owners increasingly want more visibility and control of the asset, and have a desire to "dabble with some of the knobs and dials a bit more," Hofeldt said. That's where the idea for an interactive chatbot was born, utilizing more than a decade of data gathered from projects in the field.

The platform is designed to allow asset owners to diagnose issues sooner, while enabling hands-on engagement in power trading and analysis of co-location benefits of additional resources, like solar.

The goal is to create an ecosystem of apps, similar to a smartphone marketplace, that allows asset owners to easily understand and communicate with projects in the field.

Stem isn't there yet, but they're close, Hofeldt says. The platform, Athena, could soon answer questions about inverter configurations to improve solar or storage output or analyze scenarios for project sizing.

"It turns what is a very complex system into something which is more interactive," Hofeldt said. "Being able to basically interrogate and have a conversation with your infrastructure is a pretty interesting space to be in."