ABB and Roche Show How Physical AI Can Transform Laboratory Workflows
What has happened?
ABB Robotics and Roche Diagnostics have announced a global collaboration to develop robotic automation for clinical laboratories.
The first planned applications focus on two areas.
In pathology laboratories, ABB will develop systems to support the handling and organisation of slides.
In central diagnostic laboratories, autonomous mobile manipulators will be developed to move samples and materials between laboratory instruments.
The collaboration combines ABB’s robotics, automation and software experience with Roche’s diagnostic and laboratory workflow knowledge.
For ComputeGlobal’s audience, the importance of this story reaches beyond healthcare.
It shows how physical AI, mobile robotics, intralogistics and connected automation are moving into environments where accuracy, traceability, safety and operational consistency are critical.
Why this matters for businesses and operational teams
Physical AI refers to artificial intelligence used by machines that interact with the physical world.
In a laboratory, warehouse or factory, this may involve robots that can perceive their surroundings, move materials, respond to changing conditions and coordinate with other equipment.
The ABB and Roche collaboration is especially relevant because it focuses on connecting separate stages of a workflow.
Moving a sample between instruments may appear to be a simple transport task. In practice, the process can involve identification, tracking, correct orientation, safe handling, prioritisation, exception management and confirmation that the right material reached the right destination.
The same principle applies in manufacturing and logistics.
A robot may be capable of moving a box, component or sample. However, the wider operation still needs accurate data, reliable handovers, safe routes and properly integrated systems.
Automation should therefore be assessed as part of a connected process rather than as an isolated machine purchase.
Why this matters
This story matters because laboratories contain many repetitive but important movement and handling tasks.
A member of staff may need to collect samples, organise slides or transfer materials between instruments. These activities take time and must be completed accurately.
Robotics may help with some of this repetitive movement.
However, the robot still needs to know what it is carrying, where it should go and what to do when something unexpected happens.
In simple terms, the technology must fit the workflow.
A capable robot placed into a poorly mapped or disconnected process may create new problems rather than solving existing ones.
What organisations should do next
Businesses should use this development as a prompt to examine repetitive handling, movement and tracking processes across their own operations.
The starting point should not be: “Which robot should we buy?”
The better questions are:
Where are people spending time moving materials between workstations?
Where do manual handovers cause delays, errors or weak traceability?
Which tasks are repetitive but still require consistent handling?
Can existing systems provide reliable identity, location and task data?
What happens when an item is damaged, incorrectly labelled or sent to the wrong place?
Would a small pilot demonstrate practical value before wider investment?
The answers may point towards mobile robots, machine vision, digital tracking, workflow integration or process improvement.
They may also show that the operation needs better data or clearer processes before robotics is appropriate.
Impact on different organisations
SMEs
Small businesses should not assume advanced robotics is only relevant to major laboratories or manufacturers.
The immediate opportunity may be smaller and simpler: digital tracking, barcode scanning, automated inspection or a focused material-movement pilot.
A readiness review can help determine whether the process is sufficiently stable and whether the likely benefits justify the cost.
Medium businesses
Medium-sized organisations may have enough volume to make automation commercially relevant but may still rely on manual handovers between departments.
They should examine where movement, waiting time, inspection or administrative recording creates bottlenecks.
Integration with existing software should form part of the project from the beginning.
Large businesses
Large organisations may have more opportunities to automate sample, component or material movement across multiple departments.
The main challenge is often not the individual robot but governance, interoperability, cybersecurity, maintenance and consistent deployment across sites.
Multinationals
Multinationals need repeatable automation models that can operate across different facilities, regulations and working practices.
Data standards, validation, lifecycle support and supplier interoperability become particularly important when technology is scaled internationally.
Public sector
Public-sector organisations may apply the same principles in healthcare laboratories, stores, estates, maintenance operations, public infrastructure and asset management.
Projects should be supported by clear governance, value-for-money evidence, safety assessments and appropriate human oversight.
Contractors and subcontractors
Contractors and subcontractors may use automation and tracking to improve material control, inspection records, service evidence and operational consistency.
They should also clarify responsibility for system maintenance, data ownership, incident management and integration with client platforms.
Practical automation-readiness checklist
- Define the operational problem before selecting technology.
- Map the complete workflow, including handovers and exceptions.
- Measure current waiting time, movement, errors, rework and manual effort.
- Identify the materials, samples, products or components involved.
- Check labelling, identification and traceability standards.
- Review whether data accurately reflects the physical process.
- Examine routes, access, floor conditions, congestion and safety risks.
- Assess integration with laboratory, warehouse, production, ERP or reporting systems.
- Identify regulatory, quality, validation and cybersecurity requirements.
- Decide where human approval or intervention must remain.
- Define measurable pilot outcomes before deployment.
- Build an evidence-led and, where appropriate, grant-ready business case.
Where ComputeGlobal fits in
ComputeGlobal supports organisations assessing practical applications of robotics, physical AI, machine vision and connected automation.
This may include automation-readiness assessments, workflow mapping, robotics feasibility reviews, intralogistics studies, machine-vision opportunity assessments, operational performance analysis and grant-ready automation business cases.
The purpose is not to recommend robotics automatically.
It is to determine whether the workflow, data, systems, environment and commercial case are sufficiently ready to support a responsible investment decision.
FAQ
What is physical AI?
Physical AI is artificial intelligence used in machines that sense, move and operate in the physical world. Examples include robots, autonomous mobile robots, machine-vision systems and automated inspection equipment.
What is an autonomous mobile manipulator?
It combines a mobile robotic base with a robotic arm or handling system. This allows it to travel between locations and interact with objects or equipment.
Why is laboratory automation relevant to manufacturers and warehouses?
The same core principles apply: accurate identification, safe handling, traceability, connected systems and reliable movement between stages of a workflow.
Will robots replace laboratory professionals?
The announced applications focus on repetitive handling and movement. Roche’s own guidance emphasises that AI should complement professional expertise and retain appropriate human oversight.
Can smaller organisations use this type of automation?
Potentially, but the appropriate starting point may be a smaller tracking, machine-vision or material-handling project rather than a complex robotic deployment.
What should an organisation assess before investing?
It should assess workflow suitability, data quality, safety, integration, validation, cybersecurity, workforce readiness, maintenance and measurable operational value.
Can ComputeGlobal support a grant-ready automation business case?
ComputeGlobal can help assess the operational problem, evidence base, technical options, implementation risks and expected value needed to prepare an investment or grant-ready business case.
