Will AI and robots replace 3D scanning operators and surveyors? What 9 practitioners say
Built from 9 2DIAMR conversations · updated 12 Oct 2026
Short answer
Not the people – but the tasks, yes. The guests agree that dangerous, repetitive and low-effort capture work is being automated first: autonomous drones in mines, drone docks for repeat flights, AI that plans flights or extracts features. What stays human is judgement: planning the mission, checking whether data is good, signing off legal work and deciding what to do with the results. Most guests expect more data and the same or more people, with roles shifting from carrying a scanner to managing systems and quality. They disagree on how fast it happens and on whether entry-level, button-pushing jobs will survive.
At a glance
Capture
Being automated now (per guests): Dangerous and hard-to-reach scanning, repetitive drone flights from docks, self-planned flight paths
Staying human (per guests): One-off jobs, intervening when things go wrong, deciding where and when to capture
Processing
Being automated now (per guests): Calculations, feature and crack detection, running processing pipelines with AI agents
Staying human (per guests): Checking results, knowing good data from bad, QA/QC
Responsibility
Being automated now (per guests): –
Staying human (per guests): Signing off licensed and legal survey work; governing whether data is fit for use
The job itself
Being automated now (per guests): Carrying a tripod scanner around a whole site (per Ben Williams)
Staying human (per guests): Manager of systems and data collection, targeted site visits, decisions
What is actually changing
Ben Williams of Exyn Technologies separates automation – a routine a human defines and a machine repeats – from autonomy, where you give a goal and the system works out the steps. At the highest level Exyn works at, the operator no longer flies the drone at all, but can still watch and step in.
Others see the same shift elsewhere. Dylan Gorman of PilotByte describes a drone that built its own point cloud and planned its capture flight, cutting a job of about six hours to one hour and twenty minutes. Kelly Watt says drone-in-a-box systems arrived rapidly. Rami Tamimi expects AI to extract features like curbs and guardrails from mobile mapping data, and Bilawal Sidhu already lets AI agents run a Gaussian splatting pipeline with saved settings. Arek Szadkowski of Esri expects AI and semantic search in Gaussian splats – asking a language model to find and map assets – to change how mapping is done within two to three years.
“Level four is where our systems operate, and that's where the operator is not flying the system at all.”
“All I had to do was just babysit the drone to make sure it wasn't gonna crash into anything.”
“Anything that you can come up with a standard operating procedure for, like something that you'd record like a Loom video to explain somebody else to do it, you could probably get an agent to do it.”
“Within the next two to three years we will see a huge shift in the way the mapping is done in the industry.”
Which tasks go first
Dangerous work goes first, says Ben Williams: Exyn’s first use cases pulled surveyors back from post-blast stopes where people died every year. Next come repetitive work, hard-to-access spots and monitoring jobs that weren’t economic before. Dylan Gorman and Kelly Watt agree on the repetitive part – docks are made for repeat tasks, and robots take over the boring work humans aren’t good at.
The guests disagree on whose job is most exposed. Rami Tamimi says AI hurts the labour layer most – people who push buttons and rely on someone else to check their data. Ben Williams sees the fastest adoption where highly trained people spend time on rote capture, and says replacing low-wage labour often doesn’t make sense.
“What I'll call the low-hanging fruit here is all of the most dangerous elements of people's operations, those are the places where autonomous systems go first.”
“Drone docks are made for repetitive tasks, right? And that's where the automation is really gonna come into play. If something is not a repetitive task, you're still gonna have to have an operator go out.”
“It hurts that population of folks that are basically laboring.”
Which roles stay – and grow
Rami Tamimi sees AI complementing surveyors, because a licensed person still has to check and sign off – and won’t sign what they haven’t understood. Ben Williams expects the tripod-carrying field role to go away, while the same person moves to targeted site visits and to managing a fleet of systems; the cognitive load moves somewhere else.
Kelly Watt adds governance: someone has to decide whether the data is fit for use, and robots themselves need people to set them up and manage them. Rami Tamimi’s hiring test says it all – a field person who knows what good and bad data looks like.
“The human element is still gonna be there 'cause we need to check it and make sure that it's correct and sign off on it, and we're not gonna sign off on something that we haven't looked at or understood.”
“It removes the direct link of one operator to one system, and instead becomes a manager of systems, a manager of data collection, basically.”
“You still produce data, and someone's gotta deal with the data.”
The workforce problem
Automation arrives in an industry that is short of people. Per Rami Tamimi, the average professional surveyor in the US is 57, and the hard part is the handover between generations. Ben Williams says finding people is one of the biggest challenges in Exyn’s markets, so autonomy lets one team cover what used to need three. Kelly Watt and Barry Bassnett describe experienced people leaving with knowledge younger staff don’t yet have.
Meanwhile the tools are being built for non-specialists – Dylan Gorman says many new drone-mapping people come from other industries – which makes knowing the fundamentals more valuable, not less.
“The average age of a professional surveyor, at least in the United States, is 57 years old.”
“We can't find anybody, so we need this team that exists here to be able to cover what would've been three teams before.”
Scepticism and limits
Dylan Gorman says much of what is sold as AI is automation that was always there under a new label, and that a generated 3D model can’t replace capturing a real bridge. Zlata Karpushina of FJ Dynamics calls the “magic button” one of the biggest delusions in technology, and Ben Williams admits “autonomous” is overused – customers buy outcomes, not buzzwords.
Trust is the deeper issue. Barry Bassnett wants every AI-regenerated part of a model marked, because reality capture is about capturing reality. Rami Tamimi asks whether surveyors still do QA/QC or just depend on AI. No guest addressed who is liable when an autonomous system or AI model gets it wrong.
“We just slap AI onto everything, when in reality it's just a procedure that it's running through.”
“If I'm using AI to regenerate a model, I need to know... I'm happy to take it, but I need to know which bits were regenerated.”
“A customer isn't gonna buy something just because it's autonomous, 'cause it sounds nice.”
How fast? The guests disagree
Kelly Watt sees it happening already, and Olaf Freier of INTERGEO says trends that used to take five to ten years now take about one. Dylan Gorman expects it eventually, but not right now – batteries, distances and line-of-sight rules still need people – and Bilawal Sidhu thinks we underestimate how long technology takes to spread, while warning that entry-level workers will have a hard time.
Ben Williams’s long view: even 10 to 20 years out, with autonomous systems capturing continuously, people will still go to the field. Zlata Karpushina agrees robots may gradually take over field scanning. Kelly Watt and Barry Bassnett expect more data and more work, not fewer people.
“I see a lot more data being captured. I see no fewer people in the industry.”
“Is it gonna change? Sure. But not right now.”
“It's not lasting five or 10 years. It's there maybe one year and then it smashes the floor.”
“I think that's what gradually might be happening.”
What to do now
Rami Tamimi’s advice is to move from pushing buttons to quality control – you don’t have to become a licensed surveyor, but you do need to understand what you’re doing and the fundamentals behind the equipment. Dylan Gorman says the skill level of drone pilots will have to go up, and Ben Williams says operators must know when to use an autonomous system and when not to.
Kelly Watt expects the people who embrace and harness the technology to be the big winners, and Bilawal Sidhu suggests automating your own repetitive processing with agents now – while keeping your own thinking sharp.
“Maybe you don't have to become a professional surveyor, but at least be able to do quality control and quality assurance work so that you can preserve your worth.”
“Those that embrace the technology, learn the technology, and can harness the technology will be huge winners.”
“Maybe the skill level of a drone pilot is gonna have to go up in order to maintain your job.”
Questions
- Will AI replace land surveyors?
- Rami Tamimi expects AI to complement surveyors rather than replace them, because a licensed professional must still check and sign off. The roles most at risk, per Rami Tamimi, are button-pushing field jobs without QA/QC skills.
- Will autonomous drones and robots replace scanning operators?
- Ben Williams (Exyn) expects the tripod-carrying field role to go away while the person becomes a manager of systems; Dylan Gorman says docks suit repetitive tasks but one-off jobs still need an operator.
- Which scanning jobs are automated first?
- Per Ben Williams, the most dangerous ones – such as scanning post-blast mine stopes – followed by repetitive and hard-to-access work; Dylan Gorman and Kelly Watt also point to repetitive drone flights and inspections.
- What skills should a scan operator learn to stay relevant?
- Quality control and the fundamentals behind the equipment (Rami Tamimi), knowing when to use an autonomous system (Ben Williams), and using automation and AI for your own repetitive work (Kelly Watt, Bilawal Sidhu).
Every statement above comes from what guests said on 2DIAMR, attributed to them – these are their experiences and opinions, not 2DIAMR’s own measurements.
