Development teams are being asked to screen more sites and bring more detail into early design, without any extra hours in the day. So every hour of focus counts.
Then it’s 8 AM, and the day fills up with everything except the work that matters. A colleague needs every object in your most mature project renamed for next week’s due diligence. The town hall is on Tuesday, and your design still needs trees, access roads and batteries that look real. Your boss has sent a set of wind turbine coordinates to rebuild by hand. And a red warning in your hybrid analysis won’t go away.
None of it is hard. It just takes the time and headspace you’d rather spend on land, permits and grid.
That’s the work Glint AI is for.
Developers use Glint Solar from the first land search all the way to due diligence, and that rests on two things. The first is the data. Glint Solar has built-in layers across 35 countries, with 480+ in the US, 350+ in Germany, 200+ in Italy and 150+ each in France and the UK. They cover grid infrastructure lie substations and capacity, regulations, environmental and archaeological constraints, flood risk, industrial sites, existing projects and more. The second is a set of deterministic design and analysis tools that turn that data into sites and layouts you can defend.
Glint AI sits on top of both and gives you a new way to use them. Most of us already talk to our software in plain language, and your development platform should work the same way. Say what you want, and Glint AI does it with the data and tools you already rely on.
Some design tasks are simple but painfully slow. Think of placing trees one by one along a boundary, restyling twenty exclusion zones or resizing every object in a row. They take a lot of clicks and add nothing to the quality of the design.
With the Design Agent, you describe the change and Glint AI makes it on the map. Here’s what that sounds like.
When a request could mean more than one thing, it asks first. Ask for oak trees along your site boundary and it checks whether they should sit just outside the boundary, just inside it or exactly on it. Pick a suggested reply and the trees go in. Then it offers next steps, like adjusting their height or adding a fence, and “Make all the trees 8 metres tall” resizes the whole row at once.
Glint AI runs energy analyses for you too. Ask it to “Run the analysis with tilt 10°, 15°, 20°, 25° and 30°” and it takes care of the loop. It confirms which configuration to start from, changes that one setting for each run and names every analysis after its value. Or ask it to run one analysis for each of your saved PV configurations.
Then “Compare all analyses in a table” puts capacity, specific yield and other headline figures side by side in the chat. You can also pull P50 annual energy, the loss breakdown or BESS income and cycles from a single report without opening it.
Everyone gets stuck sometimes, and you shouldn’t have to dig through documentation or wait for a support reply to move on.
Glint AI knows which project you have open, and whatever you’ve selected on the map comes along with your next message. So you can say “make this 50 m wider” or “why won’t these batteries work here?” without explaining anything first, the way you would with a colleague sitting next to you.
It answers from Glint Solar’s help content and the actual state of your project, so “How do I export my layout?” gets you steps that fit what’s on your screen.
It can’t help if you haven’t had your coffee yet, though. That one’s on you.
We made one deliberate choice here. You talk to Glint AI in natural language, but it doesn’t use generative AI to draw on your map.
Every action runs through Glint Solar’s own tools, the same ones you use yourself. When it places a BESS area, it uses the battery placement tool. When it changes a configuration, it only sets the values you asked for, and if one is missing it asks you rather than guessing. Its drawing changes go into the same undo history as yours, so Cmd+Z or Ctrl+Z works as it always has. What you get is what Glint Solar would produce if you’d clicked it yourself, just faster. And your project data isn’t used to train AI models.
Glint AI works best with one clear task at a time, so it won’t build a whole site from a single prompt. It compares technical results only and has no view of LCOE or IRR, which means the investment call stays with you. And it doesn’t screen land, since that’s what Land Screening does today. The full list of what it can and can’t do is in the knowledge base article.
The Design Agent is the first place you’ll meet Glint AI, but it won’t be the last. Our direction stays the same: turn the data in Glint Solar into better decisions, faster. Glint AI comes along for each step.
Your expertise is the scarce resource, and we’d rather you spend it on the decisions only you can make. Glint AI will handle the rest.
Glint AI is currently available to selected customers. Book a demo and we’ll run it on one of your sites.