Periodic Reporting for period 1 - tilkal (Next Generation Blockchain-based Supply Chain Traceability & Transparency Platform)
Reporting period: 2023-01-01 to 2023-12-31
With the need for resilience, the demand for transparency and the increase in regulatory obligations over the product life cycle, it has become crucial for companies to ensure that their claims and operations rely on a measure of what really happens across their chain, to drive effective and sustainable actions.
Supply chains serve as the foundation of businesses, yet they remain fragile, complex and fragmented, hindering effective communication among participants. The involvement of numerous stakeholders, often challenging to identify, exacerbates the issue, compounded by a dearth of consistent supply chain data. Consequently, it has become imperative for companies to build a solid, holistic, and transparent dialogue between upstream and downstream supply chain stakeholders that will allow to collect and aggregate operational, environmental and social information.
End-to-end, real-time traceability is the key to operating this paradigm shift and is becoming a new form of “license to operate”. Tilkal brings the traceability, transparency and auditability capabilities to answer that challenge and help businesses. The EIC project is focused on developing and piloting AI algorithms for risk analysis in supply chains, in relation with the regulatory context.
Supply chains serve as the foundation of businesses, yet they remain fragile, complex and fragmented, hindering effective communication among participants. The involvement of numerous stakeholders, often challenging to identify, exacerbates the issue, compounded by a dearth of consistent supply chain data. Consequently, it has become imperative for companies to build a solid, holistic, and transparent dialogue between upstream and downstream supply chain stakeholders that will allow to collect and aggregate operational, environmental and social information.
End-to-end, real-time traceability is the key to operating this paradigm shift and is becoming a new form of “license to operate”. Tilkal brings the traceability, transparency and auditability capabilities to answer that challenge and help businesses. The EIC project is focused on developing and piloting AI algorithms for risk analysis in supply chains, in relation with the regulatory context.
For the first year the main technical and scientific work has been focused on developing the core elements of a risk analysis and scoring model for supply chains. This resulted in two main achievements:
1) A weakness analysis model: this model uses machine learning to “understand” supply chain network topology and identified outliers in supply chain flows. Key challenges are the explainability of the results and the unsupervised context.
2) Adaptive learning foundation: Creation of a correction layer of the weakness analysis model, based on user feedback . Key challenge if the reduction of the burden on users by making intelligent selection in the data requiring feedback.
Finally a foundation for scoring end-to-end supply chain flows has been designed and implemented.
1) A weakness analysis model: this model uses machine learning to “understand” supply chain network topology and identified outliers in supply chain flows. Key challenges are the explainability of the results and the unsupervised context.
2) Adaptive learning foundation: Creation of a correction layer of the weakness analysis model, based on user feedback . Key challenge if the reduction of the burden on users by making intelligent selection in the data requiring feedback.
Finally a foundation for scoring end-to-end supply chain flows has been designed and implemented.
In the first year we have been able to go beyond the traditional reporting and scoring tools that characterize software systems that engage with complex supply chains. We have been able to leverage AI techniques to develop a system that is capable of instantly identifying anomalies in data describing a supply chain flow, based on a standard model. Moreover, this system is fully scalable - it learns from the context it is placed in and is capable of improving itself by leveraging relevant human feedback.