Industrial processes increasingly rely on complex procedures that need to be accurate, traceable, interoperable, and easily accessible. Yet, procedural knowledge is often scattered across documents, systems, or even different software tools, making it difficult to maintain, reuse, or connect with AI-based services.

Solution 5 (S5) addresses this challenge by providing the Procedural Knowledge Management System (PKMS): a unified, semantically grounded platform that stores, links, validates, and exposes procedural knowledge as a trusted, structured, machine‑readable resource.

Built on Semantic Web technologies and integrated with the new Procedural Knowledge Ontology (PKO, https://w3id.org/pko#), S5 is developed by Cefriel and acts as the core component glueing together all PERKS solutions—from extraction (S1) to authoring (S2) and validation (S3), from exploitation and reuse (S6) to auditing (S4). You can read more about other solutions in the other articles already published on the PERKS website.

Making Procedural Knowledge Interoperable and Actionable

At the heart of S5 is the goal of turning heterogeneous procedures—PDFs, notes, manuals, or tacit knowledge—into coherent, interoperable knowledge graphs.

To achieve this, S5 introduces two key contributions:

1. A reference conceptual model for procedural knowledge (PKO)

S5 relies on the newly developed Procedural Knowledge Ontology, built by reusing established Semantic Web vocabularies and driven by the requirements of PERKS industrial use cases.

PKO ensures that procedures are: consistently represented, semanticsrich and machineinterpretable, ready for validation and reasoning, aligned across all PERKS solutions.

2. A reference architecture for a Procedural Knowledge Management System (PKMS)

The PKMS integrates several components for the collection, storage, validation, retrieval, and governance of procedures.
Crucially, S5 is designed to avoid duplicating existing company data, adopt open/nonproprietary technologies, be easy to integrate with enterprise IT systems, and expose procedural knowledge through a data product accessible to any dependent application.

S5 aims to be the central knowledge hub for procedural knowledge and also offers a user interface to access other solutions and procedures.

How S5 Works

S5 implements a modular architecture that provides all essential backend services for managing procedural knowledge:

  • Storage of all procedures, metadata, constraints, and ontologies in a PKOcompliant knowledge graph, enabling versioning, validation, and semantic queries.
  • Metadata ingestion supporting a datafabric approach, linking to external repositories without duplicating documents.
  • Unified access to procedural knowledge through a data product API, built using grlc (https://grlc.io/, REST APIs from SPARQL queries)
  • A metadata catalogue offering search, filtering, and browsing of procedures through a webapp interface.
  • Automatic validation of procedures using PKO and domainspecific SHACL shapes to ensure consistency and quality.
  • Authentication, authorisation, and integration via shared Identity Provider (IdP), with external resources managed via standardized object storage API.

This approach makes S5 the central backbone of PERKS, ensuring interoperable, validated, and easily accessible procedural knowledge across solutions and use cases.

PKMS in Use Cases: Accessible, Reliable Procedural Knowledge for Every User

S5 gives users a unified and trusted environment where procedural knowledge becomes easy to find, understand, validate, and reuse. Instead of navigating scattered documents or inconsistent versions, operators, engineers, and managers can rely on S5 to quickly access the information they need through an intuitive catalogue or through applications that consume the S5 data product. Thanks to the underlying knowledge graph, validation engine, and the introduction of RoleBased Access Control (RBAC) powered by Keycloak, users always work with accurate and approved procedures—ensuring both correctness and proper authorization. This makes S5 a valuable everyday tool, supporting operational work, decisionmaking, training, and integration with custom company systems or AIenabled applications.

What Users Can Do with S5

  • Search and filter procedures based on standard metadata (titles, descriptions, status, etc.) and use case-specific fields (manufacturer, machine type, etc.)
  • Inspect procedure metadata and data, including linked resources and provenance information
  • Access structured procedural knowledge, represented consistently according to PKO (steps, tools, etc.).
  • Access linked external resources (documents, multimedia, specifications)
  • Integrate S5 data into custom tools such as dashboards, mobile apps, AR systems, or internal company platforms using the exposed data product.
  • Rely on RBACbased access control, ensuring that users only see the procedures they are allowed to see, and that sensitive or restricted procedures are properly protected.

Why S5 Is a Key Building Block

S5 provides the foundation on which all other PERKS services operate:

  • S1 extracts draft procedures → stored and validated through S5 (read more about S1 here)
  • S2 and S3 author, edit, validate and approve procedures → persisted via S5 APIs (read more about S2 here)
  • S4 audits extraction and execution → S5 collects relevant traces for procedure lifecycles
  • S6A provides deployment support → S5 enables external integration

By relying on knowledge graphs, S5 ensures that procedural knowledge is:

  • interoperable across systems,
  • reusable for future applications,
  • discoverable through metadata and APIs,
  • trustworthy and validated,
  • accessible to both humans and machines.

Importantly, S5 also prepares the ground for advanced AI scenarios: LLMs and other AI systems can benefit from structured, authoritative, ontologyaligned data, improving accuracy and reliability.

Learn More on S5

  • Learn more about how S5 technologies can support Procedural Knowledge management in industry by reading the following paper: Irene Celino Scrocca, Valentina Anita Carriero, Antonia Azzini, Ilaria Baroni and Mario Scrocca: “Procedural knowledge management in Industry 5.0: Challenges and opportunities for knowledge graphs“, in Journal of Web Semantics 84, 2025, available at https://doi.org/10.1016/j.websem.2024.100850
  • Discover PKO at https://github.com/perks-project/pk-ontology and by reading the following paper: Carriero V., et al. (2025): “Procedural Knowledge Ontology (PKO)” in The Semantic Web. ESWC 2025, available at https://arxiv.org/abs/2503.20634
  • Read more about the first release of the PKMS in the PERKS deliverable: Scrocca, M. et al. (2024): “D4.2: Tools and technologies for Procedural Knowledge Management, first release”, on Zenodo. https://doi.org/10.5281/zenodo.18340820.

Cefriel Research Team

The PERKS Solution S5 has been designed and developed by the Cefriel research team: Mario Scrocca, Valentina Carriero, Antonia Azzini, Ilaria Baroni, and Irene Celino.