We’ve detected that you are using an outdated browser. This will prevent you from accessing certain features. Update browser

Abstract data representation

Safety Accelerator startup graduate commercialises solution with HSE.

London-based startup Ohalo has signed two-year contract with GB’s Health and Safety Executive, applying their cutting-edge safetytech to anonymise and desensitise health and safety data.

LR spoke with our Safety Accelerator participants Steven Naylor, Senior Scientist in HSE's Science Division and Technical Lead on its Discovering Safety Research Programme and Kyle DuPont, CEO of UK-based startup Ohalo, about their journey with the programme, as they continue to scale and commercialise post-pilot.

How did you both become involved in the Lloyd’s Register Safety Accelerator?


Back in 2019, the HSE put in an industrial challenge to LR’s Safety Accelerator centred around the challenges of auto-redacting large volumes of GDPR sensitive research datasets. The principle driver was the aspiration to promote the sharing of routine health and safety data across the global health and safety community to stimulate innovation in use of data and the major barrier that the risk of breaching data protection legislation posed to being able to do so. Ohalo’s technology solution pitch to use their proprietary software platform, Data X-Ray, as a starting point, subsequently won the challenge. A pilot project was then run later in 2019 which successfully demonstrated proof of concept.


Ohalo was approached by one of their existing investors, Plug and Play, who was working with Lloyd’s Register’s Safety Accelerator in selecting a vendor that could meet the challenge of desensitising safety records at scale. This would enable those safety records to be mined by third parties (site operators, manufacturers, etc.) to identify safety related issues at work to reduce accidents and fatalities and get people safely back to their families at the end of the workday. 

What would you say are the biggest learnings from your experience?


The power of collaboration -  just how much you can achieve from a relatively small initial investment when you’re able to work with the right specialists, and the other avenues that you end up exploring when inquisitive minds are brought  together to work collaboratively in this way.

The work we’ve undertaken beyond the initial pilot has highlighted the value of getting as many of the key stakeholders involved at the earliest stages. Think beyond the pilot, assume it’s going to be a success and try and try to envisage the likely future direction of travel at the earliest stage and start selling the benefits to be realised as early as possible.


It really comes down to the quality of people that you work with and that has been my most positive experience throughout this both on the LR and HSE side. The willingness to try new things and iterate on what worked and what did not is so key to achieving innovative outcomes when pushing the boundaries of what is possible.
As a new programme, the LR Safety Accelerator is already achieving much more than many more mature accelerators. However, safetytech is something that benefits so many companies around the ecosystem, and I think in the future working more with LR on the business side would help drive even more impactful results.

Digital data abstract

In your opinion, why is the anonymisation and desensitisation of HSE data so important for safety and risk globally?


Many organisations are often already rather reticent in being open with their corporate datasets for reasons of losing competitive advantage, for some the perceived business risks of doing so outweigh the perceived benefits potentially realisable. Developing an effective technological solution to the problem of auto-redacting large volumes of data means the business case in being more open with data is much easier to make to the people that need persuading. 


On a technical level, natural language processing has until recently been an academic pursuit. The application of such techniques to real-world use cases that save lives sets a new benchmark and shows what the forefront of natural language processing can do in commercial applications.

How has LR and the Safety Accelerator helped you successfully collaborate and go on to commercialise?


The number of clever small start-up technology companies working in the data science and data analytics space is huge and grows year on year. The challenges of finding the right sort of company with the right sort of solution and the right sort of team that can work with you effectively on a pilot project is a huge challenge. LR’s Safety Accelerator and partner Plug and Play connected us with the best of the best for our particular challenge area and made the task of finding the right small tech start up to work with on our challenge really easy.


As a small start-up, working with government agencies can be difficult due to the processes in place. The LR Safety Accelerator was instrumental in providing an interface between HSE and Ohalo where experimentation became possible. The funds provided by LR in order to carry out this experiment were also very helpful to fund the product development necessary to tune the Data X-Ray to the HSE use case. But ultimately it comes down to people, and we were fortunate that all parties involved were very motivated towards achieving the project goals and that led to a willingness to continue to refine the initial experiment towards an actual commercial proposition at the end of the initial accelerator period.

Where do you see safetytech going in the next 5+ years?


Many representatives from across asset rich industries, such as aerospace, automotive, power and utilities, are strong advocates of use of AI to help in the inspection of assets for structural health monitoring purposes to overcome the challenges associated with manual inspection of assets. For assets in remote locations, offshore, for example, or dams and viaducts in rural locations, the safety gains to be realised by combining use of remote visual inspection technologies with AI, ML and video analytic techniques, to negate the need for workers to undertake such tasks, are obviously substantial.

The widespread deployment of CCTV and routine video capture of day-to-day working practices across workplaces, for example, construction sites and manufacturing spaces, is also opening up opportunities to exploit technologies to auto-detect unsafe working practices and precursors of serious accidents real time. This provides opportunities for health and safety practitioners to intervene early to avoid the more serious accidents happening.

One of the big trends we see in safetytech is the evolution from the physical to the digital space. This is evident obviously in the digitisation, sharing, and mining of safety reports for safety intelligence. The outcome of this in the coming years will inevitably be further digital collaboration and analysis of safety incidents and eventually real time intervention in accidents.

To give an example, what if a particular piece of equipment had a fatal failure during very cold weather conditions and the location of the same equipment could be combined with weather data from those locations to recommend a stop work command in real time? This is a possible future that will become reality with work like HSE and LR conduct.

Anonymising and desensitising health and safety data.

Hit enter or the arrow to search Hit enter to search

Search icon

Are you looking for?