From strong safety performance to even greater precision: AB Agri’s AI journey

By introducing continuous, objective insight from Protex AI alongside its existing safety processes, AB Agri has strengthened its ability to act on risk in a more targeted and consistent way.

AB Agri used the Protex AI system to introduce programmable rules for computer vision detections, establishing a consistent and objective approach to identifying Safety Learning Opportunities (SLOs). The result was a factual heat map of learning opportunities, supporting more informed prioritization and targeted improvement. The business was then able to present very clear cases for proactive capital spend and enhancing internal processes where lagging indicators were already very good.

The pre- and post-Protex AI data sets show unambiguous results:

• 91% decrease in defined WPT learning opportunities.
• 99% reduction in measured PPE non-compliance incidents
• 29% decrease in defined learning opportunities in outdoor areas
• Weekly dashboard and performance trending against all metrics

AB Agri has long prioritized health, safety and employee wellbeing, underpinned by established reporting processes, continuous improvement programmes and strong engagement with frontline teams across its sites.

As the business continues to evolve across multiple locations and increasingly complex operating environments, leaders identified an opportunity to further strengthen these foundations with more consistent, objective insight initially focused on workplace transport arrangements, where activity is dynamic and traditional observation and reporting can be subjective.

“AB Agri operate a Safety Learning Opportunity process which is distinct to near misses, where colleagues report observations or experiences to support our Continuous Improvement culture. Protex AI was able to confirm this process was working and enabled a rich overlay of objective data when targeted in these areas”, said Dave Roberts, SHEQ Director, AB Agri UK (Mono). “The system’s ability to continuously capture data on interactions and deliver a holistic view of actual daily operations removes subjectivity, including the ability to understand how interactions and movement patterns evolve over time–not just what is observed in isolated moments.”

Complementing existing reporting and engagement processes with Protex AI, AB Agri built a program of targeted and measurable improvements to proactively drive safety forwards.

AB Agri deployed Protex AI using its existing CCTV infrastructure to provide an additional layer of continuous, objective insight alongside its established safety reporting and CI processes.

“In continuous improvement, we often talk about the difference between ‘work as imagined’ and ‘work as done’,” Dave Roberts explained. “This technology gives us a clearer, more consistent view of everyday working patterns, particularly in busy areas.”

Rather than replacing traditional reporting, Protex AI complemented it, capturing routine interactions and subtle variations in behaviour that were difficult to observe or record consistently through manual processes alone. “Manual reporting will always vary depending on context and workload,” said CI Lead Jennifer Brown. “What this system gives us is a reliable, objective data set alongside that human insight.”

Crucially, the increased volume of data did not add administrative burden. “We used to spend significant time entering observations into our safety systems and then extracting and analysing the data,” Jenny added. “Now, I can request a targeted summary, such as observations from a specific camera over a month, and receive it in seconds. That frees up the team to focus on understanding the risk and implementing improvements.”

At AB Agri’s pilot site, workplace transport arrangements were a key area of focus, particularly during peak operational periods. Using Protex AI’s path-mapping feature, the team was able to visualize movement patterns across the yard over time. This provided a clearer understanding of how routes and behaviours aligned to real operating conditions, supporting more informed decisions on layout and control measures.

A set of rules were introduced in the AI logic for these paths which then recorded improvement opportunities as they were observed, continuously and in real time. Armed with this insight, AB Agri were able to present a clear capital expenditure case to deliver improved layouts and additional equipment within the operation.

The data also presented opportunities to further improve processes to manage third parties on site. Sharing anonymized video insights with these third parties enabled factual, constructive conversations and rapid buy-in to the changes made.

AB Agri’s PPE standards were well established and widely understood across sites. Through the use of Protex AI, this became a measurable compliance metric, continuously monitored and completely objective.  The team identified a specific scenario where otherwise well-intentioned behaviour could introduce unintended risk, particularly where practical adaptations to the task could gradually shift behaviour away from the intended standard. In one example, employees wore dark boiler suits over high-visibility clothing to protect it from wear or contamination during certain tasks.

“There is always a logic behind why people work the way they do,” said Jennifer Brown. “In this case, people weren’t disregarding safety requirements–they were trying to look after their equipment and work practically in challenging conditions.”

By using anonymized video clips as part of routine safety conversations and coaching sessions, teams were able to clearly visualise risk and discuss practical alternatives, making issues tangible and easy to address. A key benefit has been a sustained 100% PPE compliance, achieved by detections and notifications from the Protex AI system and vastly reducing the reliance on management oversight.

AB Agri approaches safety as a core part of operational excellence, focused on proactive risk reduction rather than reactive cost management. Alongside protecting people, leaders also look for clear evidence to ensure investment is directed where it has the greatest preventative impact.

Following strong early results, AB Agri is extending the use of Protex AI across additional factories and mills, to complement existing safety processes and support greater consistency in how insights are identified and acted upon.

From the outset, leadership has been clear that any expansion must reinforce, not erode, the trust established with frontline teams. The purpose of the technology is to protect people and work collaboratively to remove risk, rather than to introduce additional controls or oversight.

• Be transparent from the start: Explain how the technology works, and how the insights will be used to prevent harm, not to evaluate individuals.
• Keep the purpose human-centred: Position the system as a tool to remove hazards and make jobs safer and easier, especially in complex areas.
• Involve frontline voices: Engage employees and supervisors early so concerns can be addressed and practical improvements shaped by those doing the work.
• Introduce it where collaboration is strongest: Start in sites with open communication and a mature safety culture, then share lessons learned to support other locations.