A single diving-led inspection of a North Sea pipeline can run into six figures before it even starts. A typical 50km stretch takes a crew of six to eight divers ten to 15 days and costs $250,000–400,000. Prices climb further if the structure sits in deep water and requires saturation diving equipment.
Subsea inspections also involve significant physical risk and the task remains among the most hazardous jobs in the industry.
Go deeper with GlobalData
Discover B2B Marketing That Performs
Combine business intelligence and editorial excellence to reach engaged professionals across 36 leading media platforms.
However, across the oil and gas value chain, robots are starting to change who, or what, assumes the risk, as three presentations at the Oil and Gas Automation and Digitalisation Congress (Automa) 2026 will address. The conference takes place in October in Amsterdam, the Netherlands.
Upstream: replacing the diver’s black box
Francis Wanmi Ticky, a robotics engineer at Saudi Aramco, will present on a mini remotely operated vehicle (ROV) trialled for subsea structure inspection at an offshore platform.
Conventional inspection requires specialised divers to run visual, ultrasonic and cathodic protection tests, before returning with photographs and sensor readings – a process that is costly, hazardous and hard to independently verify.
The ROV evaluated by Saudi Aramco’s team carries a high-definition camera, alongside ultrasonic and cathodic protection probes, plus cleaning equipment to clear marine growth before it takes a reading. Its results matched previous diver-led inspections closely enough to build confidence in the solution, but the bigger shift is in the level of insight offered by the mission. Live video is continuously streamed, and AI analyses defects and surface discontinuities as they are found, with real-time oversight by inspectors on the call, rather than reconstructed afterward from a diver’s report.
Midstream: sensors that move instead of multiplying
Max Leon Fritsche, a robotics engineer at Security Robotics, will speak on a different version of the same access problem, across pipelines and midstream infrastructure.
Installing fixed sensors throughout an entire site is often too expensive and inflexible to justify, and there are limited available skilled personnel to perform inspections. Fritsche will suggest that autonomous robots, quadrupeds, wheeled platforms and drones should act as mobile sensor carriers instead: reading analogue gauges, spotting heat anomalies, monitoring restricted areas and repeating the same round as often as needed.
What matters most for scaling this beyond a single demo site is a hardware-agnostic fleet, allowing operators to mix whichever platform suits a given task, while keeping one consistent workflow for mission planning and data analysis behind all of them. Fritsche will tell Automa attendees that robots can take on repetitive rounds and assume risk, so technical staff can spend their time on diagnosis and decisions instead, rather than replacing the judgement call itself.
Downstream: watching a site before something goes wrong
Elisa Paiano, from Saipem’s health, safety, environment and quality research and development and AI catalyst team, brings the same logic to a downstream construction site. She will consider the build phase of a new biorefinery, where workers, moving equipment, lifting activity and shifting work fronts create risk conditions that change faster than manual supervision alone can always track.
Saipem’s platform runs live video analytics against detection rules built for specific risk scenarios, flagging potential deviations, personal protective equipment non-compliance, unsafe proximity to machinery and work under suspended loads. These appear on a supervisor’s dashboard in near real time.
Final judgement remains with a person through a validation loop: supervisors mark each flagged event as a true or false positive, and that feedback retrains the model against the site’s actual layout and behaviour patterns over time.
Paiano’s own framing will be clear about the limits of the technology alone: AI doesn’t make a construction site safer by itself, it makes the site’s existing safety processes and field leadership more consistently applied.
The economics behind the shift
Taken together, these three Automa presentations will demonstrate the diversity of applications for robots in oil and gas, which improve efficiency and assume risk. A subsea ROV, a quadruped inspection fleet and a vision-AI safety platform have little in common mechanically but are connected by the value chain.
The economics behind that shift are hard to ignore. Unplanned downtime alone cost the world’s largest industrial companies an estimated $1.4tn a year in 2024, up 62% since 2019. Inspection sits at the intersection between cost and safety: catching a problem early depends entirely on how often, and how safely, someone looks.
That thread will run across the much wider Automa 2026 programme, covering the full breadth of oil and gas operations.
