Our research is structured around three key science problems:  

• How can we design AI systems to ensure trustworthy, compliant and adaptable processes and supply chains.   

• How can we design adaptive AI-based systems that preventatively mitigate supply chain risks.  

• How can we design AI systems that optimise processes and boost productivity, while fulfilling (inter)national customer and trade requirements. 

These science areas result in technology that provides the following core capabilities: 

• Modelling, optimisation and generation of processes, to provide efficient and agile support to changing business environments; 

• Techniques for automating business process compliance, to ensure provably compliant processes by design and at runtime; 

• Process analysis, process mining and prediction technology; 

• Digitisation of regulatory requirements for machine-supported analysis. 

Organisational processes are backed by information systems, which digitally support the tasks and their data requirements and storage. In information systems, processes and data are dependent on each other and cannot be designed in isolation. Both are commonly modelled using graphical tools, which assist in visualisation as well as verification of their correctness. Process models show the explicit normative behaviour of organisational processes, specifying the allowed and required tasks and activities to achieve a particular business goal. Data models, on the other hand, specify what data is used and how it links together. Both processes and their data are subject to various requirements, which are imposed by e.g., regulations, business rules or guidelines. Compliance to such requirements is either hard-coded in the systems themselves or is only implicitly incorporated and subject to after-the-fact auditing. In our research, we create an explicit digital representation of these requirements, such that they are both human understandable and machine interpretable. As a result, we can integrate all three aspects, to ensure correctness of processes and compliance with regulations, standards, guidelines and business rules.

Our sector-agnostic approach to digitising requirements has the following key advantages: 

• Requirements are neither hard-wired nor hard-coded: rules are understandable and flexible. 

• Our digital format allows a 1-on-1 relation between regulations and digitised rules, supporting exceptions, exemptions and amendments through superiority relations between rules. 

• There is an explicit mapping between policy-based regulations and lower-level quantifiable methods for measuring compliance. 

Verifying compliance of a business process by design against a given set of requirements guarantees that when executed, the outcomes on all pathways always satisfy the given requirements. This also allows to query requirements and helps optimise processes to improve performance without compromising compliance. 

DAMOCLES™ can monitor the ongoing execution of a business processes to determine whether such execution aligns with the expected behaviour and subsequently identify and report when that is not the case. This ensures continuous and objective evidence of compliance during execution, while supporting changeability of running processes. 

When unexpected behaviours occur during the execution of a business process, DAMOCLES™ can be used to forecast the continuation of the execution resulting from the unexpected behaviours, allowing to pre-emptively identify possible scenarios where continuation of the process can lead to compliance risks. 

After identifying executions of a business process that produce breaches against the requirements, DAMOCLES™ can be used to identify possible adjustments to the running process instances and prevent these breaches from occurring. 

The team was involved in a project funded by DAFF National Agriculture Traceability Regulatory Technology Research and Insights Grant Round to establish a data-led system that enables continuous assurance of red-meat processing. The prototype aimed to reduce the burden associated with achieving regulatory compliance and to provide decision-support and tracing capabilities through an interactive dashboard. This was achieved through federating data streams and mapping them for regulatory compliance pathways. Subsequently, DAMOCLES™ was used to digitise and provide live updates on regulations regarding food safety, animal welfare, export requirements and business objectives.  

The process was tracked all the way from cattle grazing on the farm, their truck-journey to the processing facility, time spent in lairage and subsequent processing and packaging by the processor. This tracking was enabled by real-time data from four types of sources; Weather, Farm, Transport and Processor. These 4 sources send live events capturing what has occurred on-the-ground, which are generated by a combination of automatic sensors and manual entries.  

Subsequently, the data-streams are federated at the Event Store and sent to the DAMOCLES™ API for live analysis based on the digitised requirements provided. Our patented Cross-Instance Compliance Checker technology effectively and efficiently achieves this by aggregating and storing necessary information given by each event, performing calculations and projections, and monitoring the status of each requirement. After receiving an input event, the DAMOCLES™ API promptly sends compliance outcomes and projections back to the event store.  

The UI retrieves information at the event store for displaying several modules; live process status, retrospective tracing, and plant compliance trends. Within these, compliance rule failures, warnings and indicators are displayed, which enables decision-support for processor management, as well as automated auditing for a regulator.  

This is a demonstration project with one of the world’s leading companies in agricultural industry. The goal of the project was to design and implement a blockchain-based event communication and tracking solution in the cotton supply chain for proprietary seeds, while monitoring the supply chain for unforeseen and deviating behaviour or events. 

The supply chain traceability platform is a blockchain-based system that collects and stores events from each stakeholder in the supply chain. Upon arrival, these events are forwarded to DAMOCLES™, where they are used to verify compliance across the supply chain. These compliance results are subsequently returned to the supply chain traceability platform for record keeping and visibility. At the same time, a live overview of the compliance status is displayed at a dashboard for the compliance manager, allowing to actively monitor compliance issues and mitigate where necessary.

The project involved multi-party data collection and sharing, to facilitate: 

• Real-time end-to-end compliance monitoring  
• Non-deniable persistent data for auditing  
• Early warning on potential compliance issues 

For example, some of the supply chain compliance events that were monitored were the following rules: 

• Planting  
  • Isolation type not approved before planting.  
  • A non-approved gin is pre-selected for cotton module processing.  
  • Equipment cleaning step missing after planting.  
• Harvesting  
  • Expected yield below threshold.  
  • Equipment cleaning step missing after harvest for stewarded cotton.  
• Ginning  
  • A stewarded cotton module appears in a non-approved gin.  
  • A module’s GPS location appears outside the field boundary.  
  • Most cotton modules were delivered, but some remain missing, e.g.,  
    • 60% of modules remain undelivered over a week.  
    • 80% of Grower’s fields are delivered but others remain missing for over 2 weeks.  
  • A module is getting ginned under different fields.  
  • A stewarded module is ginned with commercial modules.  
  • Seeds produced with at least one stewarded module are not sent to domestic use.

This project was funded through the Modernising Agricultural Trade Program – Protecting Australia’s Clean Green Brand, Department of Agriculture, Fisheries and Forestry.

The project has delivered a technology roadmap to achieve automated end-to-end export compliance. The roadmap details the requirements to embed compliance into digital supply chain systems, giving compliance activities the ability to scale quickly and remove potential bottlenecks that hinder industry aspirations to double exports.

The project has developed and trialled approaches to digitising compliance, using export protocols for red meat and horticultural as case studies. Utilising digital supply chain platforms and Australian Government systems, this helps to facilitate compliance with export regulations across the entire supply chain. This work established the foundations of developing a digital platform on machine-readable and up-to-date protocols.

A showcase of the project can be found here:
https://www.agriculture.gov.au/about/news/enhancing-agricultural-traceability