How ExxonMobil Uses Visual Positioning to Transform Field Operations | Niantic Spatial, Inc.

Energy facilities run across billions of square feet, in some of the most diverse and remote locations on earth. Finding a single valve, flange, or inspection point inside one of them still looks a lot like it did 40 years ago: paper handoffs, repeated field trips, and a worker squinting at a photo trying to figure out where it was taken.

Niantic Spatial's Eric Johnsen sat down with ExxonMobil's Michael Hotaling to talk about how visual positioning closes that gap, and why ExxonMobil is building an open Digital Reality Ecosystem to get there. Watch the full conversation below.

Hotaling has spent 35 years at ExxonMobil across operations, maintenance, and engineering roles, and now leads technology scouting through the company's venture arm.

Below is an edited version of the conversation.

Why does the energy industry need visual positioning?

Eric Johnsen, Niantic Spatial: When I say localization, I mean finding something and navigating to it accurately, down to the centimeter, with pose. You know which way you are looking, not just where you stand. How does that change field work?

Michael Hotaling, ExxonMobil: Think about how we used to travel. My parents went to the AAA auto club and ordered TripTiks, stapled cutouts of maps with a highlighter showing the route. My job in the back seat was to find mile marker 120 on the page. Planning a vacation took two weeks.

Today it is a full tank of gas, let's go. The app tells you when you will arrive, where the gas stations are, where to eat. The cognitive load drops to almost nothing.

Now walk into an industrial facility and watch how we do maintenance. A crew gets a piece of paper listing inspection points, then makes four or five trips to the field just to confirm they can reach one reading. A gate valve usually does not even exist as an asset in the ERP, not because the system cannot hold it, but because nobody can find it repeatedly without massive overhead.

The common language across all of it is visual. Everybody knows where that valve was. Visual positioning gives us a way to deliver that to anyone, anywhere, at any time.

Where does localization create real business impact?

Johnsen: You coached us toward turnarounds early on. Walk through why.

Hotaling: A turnaround is three steps. Turn off the flow of energy to keep people safe, perform the inspection or maintenance or construction, then turn the energy back on. Localization helps workers find the exact equipment, valves, and flanges they need to isolate, and confirm the status back to site managers and head office.

We have chased this for years with Wi-Fi triangulation, BLE, beacons, QR codes, RFID, ultrawideband. In an industrial facility none of it works well, and at best it tells you roughly where you are. It does not tell you what you are looking at. It cannot tell you which valve in a row of valves you actually need.

What did the field pilot reveal?

Johnsen: Any reactions from the field that stuck with you?

Hotaling: When we talk about vision at industry forums, people tell me it will be great in five or ten years. During our pilot, one of the ecosystem suppliers watched it work and said, holy crap, this is now, this is not two years from now.

There was no question of whether it was possible. For a lot of users it was the technology they had been waiting for. The work ahead is integration and scale, not validation.

Why is ExxonMobil collaborating openly with competitors?

Johnsen: You have introduced me to your direct competitors, which I do not see much in other industries. Why?

Hotaling: Each of us tried this transformation alone and ended up with customized solutions that do not talk to each other. The suppliers and contractors work for all of us. They walk into our plant to do a task, then walk into a competitor's plant to do exactly the same one. When our capabilities differ, we breed cost into the business. The competitiveness lives in how well people execute and how well people scale. So sharing what works and what does not, makes the whole industry move faster, and that is good for ExxonMobil.

Why "reality first" instead of the engineering model?

Hotaling: People challenged us constantly. Why focus on reality when the engineering model should be the center of everything? Because engineering models are built to ensure asset integrity, not to navigate. The yellow handrail, the bright red fire extinguisher, the reinforcement I-beams added after construction, those are the cues a worker actually uses, and they rarely exist in the model.

Reality first means I scan, I get real-time contextual information about the asset as it actually is, then underpin it with the engineering data.

What about robotics?

Johnsen: How do robots fit, and what has to be true to make them work in physical space?

Hotaling: Energy is behind automotive on standalone robotics, partly because our environments are huge and dispersed rather than a single structured assembly line. But the pressure of the energy transition is pushing us toward drones and quadrupeds, first to take people out of harm's way.

Across all of it, geospatial location is critical to doing things repeatedly. Computer vision and visual positioning are the lead case for doing that at scale. There will not be one robot that rules them all. The last thing the industry wants is to couple every robot with its own separate navigation system. Technology that is portable across platforms and agnostic of the hardware is what accelerates adoption.

The carrot cake principle

Hotaling: Best-of-breed functionality is like a carrot cake, 20-some layers held together with a little cream cheese, where the cream cheese is how the ecosystem plays together. The point is that when the cake is finished, I as the consumer can still afford to eat all the layers.

Early on, suppliers tried to turn my carrot cake into one big New York cheesecake, one fat layer that does everything. None of it was best in class. If you buy the biggest cheesecake, you lock out the startups and innovations that could move the needle, because they want to be one great layer, not the whole cake.

MTTR (Mean Time To Reality)

Hotaling: My colleague Kyle quotes MTTR, which most of our industry reads as Mean Time To Repair. In the Digital Reality Ecosystem, it means Mean Time To Reality. How useful is your map if the traffic jam it shows is from last week, or the gas station you pulled off for is already gone?

Our workforce demands reality as close to now as possible. We are in the infancy of driving that down, and that is where we are looking for more people to help.

Niantic Spatial's vision

At Niantic Spatial we are building a real-world foundational model for physical AI that lets humans and robots work safely and efficiently together over time. It starts with finding a valve in a refinery and expands from there. The oil and gas industry has real advantages, deep funding and clearly defined pain points, alongside its own challenges, and the same technology applies across a broad set of enterprise industries.