Stephanie Dockstader has built her career at the intersection of technology and problem-solving. As a product manager and business lead at Esri, she's spent years working across different industries with one consistent focus: listening to people's problems and bringing together the right resources to innovate solutions collaboratively.
"It's really problem solving," Dockstader explains. "And I think that's what's kept me thriving throughout my career, even as I've moved across the different industries and roles that I've had."
The Challenge of Building Digital Twins
Working with federal, state, and local governments, Dockstader has identified three major hurdles agencies face when building data foundations for reliable digital twins. The first is simply understanding what data they already possess.
"In the state and local, they have periodic collects where they're collecting over and over again. They're collaborating and sharing data at the county and the federal level, and we find just sometimes they don't even know what data they have in their own holding," she notes. Beyond inventory, agencies need to curate and organize this data in ways that make it accessible and useful.
The second challenge is developing an enterprise vision that extends beyond hardware modernization. "You need to know what you want your system to look like and do in the future," Dockstader emphasizes. "If you wanna do AI and automation, or you really wanna get your digital twins to be dynamic models that are really useful, then you need to consider imagery now as a foundational part of your system."
The third hurdle is a knowledge gap.
"There's still a big gap in just knowledge on the GIS side of what these imagery layers can do, how they can add value to the existing workflows that they have," she says.
Bridging Two Communities
Dockstader is a vocal advocate for bringing the remote sensing and GIS communities closer together—a connection she sees as critical for the future of interconnected digital twins.
"The GIS community and the imagery remote sensing community are coming together. They're overlapping more and more, but a lot of times they're talking past each other. They have a different vernacular, and they also are looking to solve slightly different problems," she explains.
Esri plays a unique role in this convergence, with expertise spanning both communities. "We have one foot in the imagery community and one foot in the GIS community. And we do a lot of work to translate between those two groups so that the technology will advance."
Her advice for technical experts struggling to communicate with decision-makers? Work backwards. "You have to stop thinking sensors and technology and computers and you have to work backwards in really trying to understand: What do I need to solve my problem? What are the key elements for the decision makers?" she advises. "Sometimes you have to work backwards to really understand what that is."
The Next Breakthrough: Geospatial Foundation Models
When asked about underappreciated breakthroughs that will have major impact in five years, Dockstader points to geospatial foundation models (GFMs) and their resulting embeddings.
"These models have been trained specifically for remote sensing imagery and data," she explains. "Most of our AI models have come from the computer vision field, and they've been trained mostly on just pictures and regular natural images and not satellite or aerial collected images - and they're different."
What makes embeddings particularly exciting is their ability to look beyond individual pixels. "Instead of just looking at the reflectance of the pixels, it's gonna look at the context of the surrounding environment. It could use lidar to tell you about the height. It can talk about soil moisture, and so you're getting all these characteristics of what that object is for every feature on the earth, but it goes across the different sensor types."
This capability addresses a fundamental challenge in digital twins: integrating diverse data sources. "By using these embeddings now, we can start tying 'em together by object instead of by pixel, but still leveraging that pixel rich data to get that information."
Looking Ahead to GEO Week
Dockstader will be among the Esri experts presenting in just a few weeks at Geo Week in Denver, where the company is contributing in six invited talks and a workshop on reality mapping. Her presentation, “Modernize Your Gis With Imagery and Remote Sensing” during the session Applied Image Processing, Analysis, and Classification will focus on a critical but often underestimated aspect of model training: data volume and diversity.
"I think people underestimate how much you need to really train a robust model and the diversity you need," she says. "You probably can't do that with your own data holdings. You're gonna need to pull in other data."
Geo Week represents an important opportunity for Esri to engage with both the ASPRS academic community and the AEC industry professionals. "It's always a good event to share our expertise out to the bigger community," Dockstader notes, adding that she's particularly focused on encouraging mid-career professionals to take on leadership roles in ASPRS.
As digital twins evolve from visual 3D representations into living, operational models, the convergence of remote sensing and GIS will become increasingly essential. With innovations like geospatial foundation models on the horizon and continued efforts to bridge technical and operational communities, the foundation is being laid for the next generation of geospatial technology.
Don’t miss sessions Global Lessons in Urban Digital Twins and Applied Image Processing, Analysis, and Classification where Stephanie Dockstader will be sharing her insight. Check out the Geo Week conference program, and get ready to join the conversation!
