For operators
Care home operators are being sold a robotic future while managing a very human present. This guide sets out which robot pilots have actually happened in UK care, what the workforce numbers really say, and how to plan without falling for hype.
Quick answer
The UK's real care home robot experience consists of companion and social robot pilots, most notably the CARESSES trial of Pepper in Advinia Health Care homes and the hygiene-tested introduction of the PARO seal with Sussex Partnership NHS Foundation Trust. No humanoid robot is in routine use in any UK care home. The workforce pressure is real, with Skills for Care reporting around 111,000 vacancies in 2024/25, but today's robots change how staff time is spent rather than how many staff are needed. Humanoids are proving themselves in warehouses first; care will follow later, in non-contact roles before anything hands-on.
Important note: this guide is general information for operators, not legal, regulatory or clinical advice.
Strip away the press releases and the documented record of robots in UK care settings is short, specific and worth knowing in detail.
The CARESSES project remains the most substantial UK care home robot trial on record. Led by the University of Genova with the University of Bedfordshire, Middlesex University and the care provider Advinia Health Care, it ran a single-blind randomised controlled trial of a culturally aware version of SoftBank's wheeled Pepper robot across care homes in England and Japan, with results published in 2020. Residents used the robot for up to 18 hours over two weeks. The research team reported significant improvements in mental wellbeing measures and reduced loneliness among participants using the fully culturally competent version. The honest caveats: it was a short trial of a conversation robot, the robots required researcher support, and Pepper itself was later discontinued from production by its maker. The trial demonstrated that social robots can engage residents; it did not demonstrate an operational product.
The PARO robotic seal's UK journey is instructive for any operator, because the hard part was not the robot. Before PARO was cleared for use on a dementia ward run by Sussex Partnership NHS Foundation Trust, the University of Brighton spent nine months testing whether a furred, handled device could meet infection control standards, as reported by Digital Health. Subsequent UK studies of PARO in inpatient dementia settings reported it was well received, while being open about small sample sizes. The lesson: in care, the deployment questions are hygiene, consent, dignity and staff workload, and they take longer than the technology.
Japan has trialled care robots more extensively than any other country, spanning companion devices such as PARO, communication robots and experimental lifting-support machines, driven by an acute demographic squeeze. The consistent finding reported across that experience is that engagement and monitoring devices find a role, while physical assistance robots have repeatedly proved harder to deploy than expected. UK operators should read the Japanese experience as a caution against assuming physical care robots are close, not as proof they are.
The commercial logic for care robots always starts with staffing, so it is worth being precise about the numbers rather than waving at a crisis.
Skills for Care, the workforce body for adult social care in England, reported approximately 131,000 vacant posts in 2023/24 and around 111,000 in 2024/25. Its most recent data shows the vacancy rate falling to its lowest level in about a decade, helped by international recruitment and pay movement, yet still running at roughly three times the vacancy rate of the wider economy. Looking forward, Skills for Care estimates the sector will need in the region of 410,000 additional posts by 2040 to keep pace with an ageing population.
Two operator conclusions follow. First, the pressure is structural and long-term, which is exactly why serious money is flowing into care robotics research. Second, nothing on the market today substitutes for a care worker. Current robots can take on engagement sessions, reminders, monitoring and some logistics, which changes how scarce staff hours are spent. That is a real but bounded benefit, and it is the correct frame for any purchase decision made this year.
| Role | Examples | Status, September 2026 |
|---|---|---|
| Engagement and companionship | PARO, companion devices, Pepper (trials) | In use and trialled in the UK; best-evidenced category |
| Monitoring and telecare | Falls detection, acoustic monitoring | Mature and widely deployed |
| In-building logistics | Delivery and cleaning robots | Established in hospitals and hospitality abroad; early in UK care |
| Humanoid non-contact tasks | Fetching, tidying, engagement | Not yet; humanoids currently proving themselves in warehouses |
| Hands-on personal care | Transfers, washing, dressing | No product exists; long-horizon research problem |
Humanoid robots are real, funded and entering work, but not in care. The reported commercial deployments are in logistics and manufacturing: Agility Robotics' Digit working in GXO Logistics warehouses under a robots-as-a-service agreement, Figure's robots at a BMW manufacturing facility, and Apptronik's Apollo piloting with Mercedes-Benz, all as reported by the companies and trade press. Even in Britain, the London-based company Humanoid has reported warehouse-style proofs of concept with Siemens rather than anything in a care setting. Our directory partner tracks this market at warehouse humanoid robots in the UK.
Why warehouses first? Because they are structured, repetitive and tolerant of error. A dropped tote is a statistic; a dropped resident is a tragedy. Care is the opposite environment on every axis: unstructured spaces, intimate contact, residents who may be confused or frightened by a machine, and a regulator, rightly, watching closely. The Care Quality Commission regulates providers rather than devices, which means any robot a home introduces sits inside the provider's own duty of care, consent processes and safeguarding responsibilities. Our guide to CQC regulation and robots in care homes covers this in depth.
A sensible planning sequence for operators, in our view: companion and engagement devices now, where resident response justifies them; building logistics and monitoring as the market matures; humanoids in non-contact roles when, and only when, they have a proven safety record in easier industries; and hands-on robotic care treated as a long-horizon prospect with no date worth planning around. Operators who want to evaluate the technology early can also consider demonstration days and events through our network partner Humanoid Robot Hire, which is a lower-risk way to see the state of the art than buying it.
For a step-by-step approach to running a small evaluation properly, see our guide to running a care home robot pilot, and for the general safety picture, Humanoid Robot UK's guide to humanoid robot safety.
Before any trial or purchase, put these to the supplier in writing. Where has this device been deployed in a UK care setting, and can we speak to that operator? What infection control testing has it been through? What happens to resident data, where is it stored, and what is your position if we terminate? What staff time does it genuinely require per week? What is the total cost over three years, including support? And what published evidence, not testimonials, supports the claimed benefits?
A good supplier answers these directly. Evasion on any of them, especially the reference site and the evidence, tells you what you need to know. Our guide on questions to ask about robots in care expands this list, and providers considering the funding side should read care robot costs and funding.
Common questions
The documented UK experience centres on companion and social robots. The CARESSES research project trialled a culturally aware version of the wheeled Pepper robot in Advinia Health Care homes in England, alongside care homes in Japan, in a randomised controlled trial whose results were published in 2020. The PARO robotic seal was hygiene-tested for nine months by the University of Brighton with Sussex Partnership NHS Foundation Trust before being cleared for ward use, and has been studied in UK dementia settings since. No walking humanoid robot is in routine use in any UK care home as of September 2026.
Not on any near-term horizon, and operators should be sceptical of anyone claiming otherwise. Skills for Care reported around 111,000 vacant posts in adult social care in England in 2024/25, and although its latest data shows the vacancy rate falling to its lowest level in about a decade, it remains roughly three times the rate of the wider economy, with an estimated 410,000 extra posts needed by 2040. The robots available today address engagement and monitoring, not personal care, so they change how staff time is spent rather than how many staff are needed. Humanoid robots capable of physical care tasks do not yet exist as products.
No credible date exists, and this site will not invent one. Humanoid robots are entering warehouses and factories first, through reported pilots by companies such as Agility Robotics, Figure and Apptronik, because those settings are structured and mistakes cost totes rather than injuries. Care settings are the opposite: unstructured, intimate and safety-critical. A sensible operator assumption is that humanoids will appear in care homes first in non-contact roles, such as logistics within the building, engagement and monitoring, before any hands-on assistance, and that regulated physical care from a humanoid remains a long-horizon prospect.