The tasks below translate the AI4Access Description of Action into a practical implementation approach for the Bio-Hub team. This is a draft of a working structure, not a replacement for the formal work packages and tasks — see the tracker for those.
The core sequencing principle
Develop the Navigator from real user needs towards the underlying knowledge and technical system, rather than starting with comprehensive data collection. The work packages need to operate as one iterative process:
User needs and queries→Knowledge requirements→Structured and contextual content→Technical implementation→Validation→Refinement and deployment
1
Define target users, use cases and expected Navigator performance
AimEstablish what we need the Navigator to do before determining the information and technical structures required to support it.
- Define the initial target audiences and the substantially different user journeys the Navigator should support.
- Compile a representative set of existing user questions and access enquiries.
- Use these to develop and test user personas and journeys, with validation through the Node Management Expert Group and other relevant stakeholders.
- Define how the Navigator should establish relevant user context during an interaction, including the distinction between anonymous discovery/exploration and authenticated/returning-user journeys using LS-AAI and available account information.
- Agree what "good enough" Navigator performance means for the main use cases and develop initial criteria against which recommendations can later be validated.
Early priorityA sufficiently representative query set and agreed core use cases should be available early enough to drive the data-model and technical discussions.
Related formal tasksWP1T1.1WP1T1.2WP1T1.3
2
Establish Node engagement and co-design
AimTreat Nodes not simply as data providers, but as requirements sources, data owners, validators, pilot environments and eventual adopters of the Navigator.
- Develop a simple and consistent project narrative for Nodes: added value, expected contribution and likely implications for Nodes and facilities.
- Prepare basic supporting material: project webpage, short presentation and FAQs/background information.
- Map Nodes according to their likely relevance for different aspects of development; initiate structured 1:1 discussions with all Node leadership.
- Use these discussions to identify existing practices, concerns, information structures and potential early pilot cases.
- Develop realistic incentives for participation and continued data maintenance, with particular emphasis on good UX, use of the Navigator on Nodes' own websites, and low-burden processes.
- Use the Node Management Expert Group and other appropriate groups for structured co-design and validation throughout development.
PrincipleNode engagement should start before the data model is fixed. A core question throughout should be: can this structure represent your services correctly, and what are we missing?
Related formal tasksWP2T2.2WP2T2.4
3
Develop the ontology-aligned data model and PID strategy
AimDefine the minimum structured information required for reliable service discovery and guidance, while ensuring that the resulting system can realistically be maintained across a distributed infrastructure.
3.1 Minimum information requirements and current landscape
- Determine the minimum information required to answer the priority user queries and support the identified personas and journeys.
- Consider both the technical minimum required by the Navigator, and information that would independently improve human-led Euro-BioImaging access support and services.
- Map existing data and its current structure — both at the Hub and in the Nodes' own webpages — plus the landscape of access/facility management systems.
- Identify gaps between the required and currently available information and assess what can realistically be collected and maintained.
- In parallel, update and rationalise technology/service terminology with the relevant Expert Groups, aligning with the FFBI ontology and an imaging-technology knowledge graph.
3.2 Access/service data model
- Develop iteratively: desk-based first model → Expert Group co-design/stress test → revision → small-scale population and technical testing → validation → network-wide collection.
- Treat updatability and distributed ownership as design requirements from the outset, rather than problems to solve after initial data collection.
3.3 Image-data model and relationship between WP2 and WP3
- Suggested working distinction: WP2 is service-provision knowledge — what Euro-BioImaging, its Nodes and relevant Hub services can provide. WP3 is the image-data journey — how imaging, data, analysis and data-management capabilities connect into viable research workflows.
- Node-based image-analysis and data-management services remain part of the common service model, with extensions where needed; external/community data resources can form a distinct but semantically connected knowledge source.
- WP3 should particularly model relationships across the imaging-data journey: acquisition method → data characteristics → analysis requirements/tools → compute requirements → data management → deposition.
- The aim is enabling conceptual guidance by the Navigator in conversation with the user across the research pipeline — not necessarily technical execution of every workflow step.
3.4 PID strategy
- Map the current PID landscape, including relevant work through foundingGIDE and community initiatives.
- Determine which entities genuinely require persistent identifiers for Navigator functionality, provenance and interoperability.
- Develop pragmatic recommendations aligned with both the emerging data model and what Nodes can realistically implement.
Overall principleKeep information in the system or community best placed to own and maintain it; create enough semantic alignment for the Navigator to combine it intelligently.
Related formal tasksWP2T2.1WP3T3.1WP3T3.3
4
Build and populate the knowledge environment
AimMove from the data model to a sufficiently rich, maintainable body of knowledge from which the Navigator can generate useful and evidence-based guidance.
- Node/service knowledge: structured information on Euro-BioImaging technologies, expertise, services, access conditions and providers.
- Evidence and contextual knowledge: existing user projects, use cases/user stories, publications, Euro-BioImaging policy documents, guidance and best-practice material needed to justify useful recommendations.
- Website knowledge: distinguish information that needs to become authoritative structured data from content that can remain unstructured but must be reliably retrievable with provenance — explore lightweight metadata and automated ingestion/indexing through the WordPress REST API, with appropriate validation and update mechanisms.
- Training knowledge: assess required changes to Microscopy.DB structures, how information can feed into the wider knowledge environment, and Euro-BioImaging's future role in the wider Microscopy.DB sustainability conversation, including a new member on its steering group.
- Funding knowledge: determine the appropriate scope, source and update model for funding information relevant to user guidance.
- Knowledge graph of the image-data journey: define how imaging technology, produced data, analysis and data-management capabilities connect into viable workflows, connect relevant resources, and identify places for intervention with users in the access journey.
- Across these sources, define a source hierarchy determining which information the Navigator should treat as authoritative and how conflicting or outdated information is handled.
PrincipleThe aim should not be to centralise all information in a single database, but to connect appropriate sources through the common semantic model.
Related formal tasksWP2T2.2WP2T2.3WP3T3.1WP3T3.3
5
Establish validation and human oversight
AimBuild validation into development rather than treating it as a final-stage check.
- Define human-in-the-loop processes for checking recommendations and correcting underlying information.
- Develop validation criteria linked to the priority use cases and "good enough" performance defined under item 1.
- Test outputs with domain experts and Node staff.
- Recruit less imaging-experienced/"naïve" users to test whether the Navigator genuinely lowers entry barriers.
- Where possible, follow real user projects through the journey from initial question to service selection/access and downstream data needs.
- Ensure findings feed back into personas, data requirements, knowledge sources and system behaviour.
Related formal tasksWP1T1.4WP6T6.2
6
Prepare integration and adoption at Nodes and facilities
AimEnsure that the Navigator can become part of Euro-BioImaging's distributed service environment rather than remaining a standalone central tool.
- Node-facing Navigator integration: approaches for embedding or linking the Navigator within Node websites and national access environments.
- Facility-management integration: map systems currently used across facilities and assess realistic opportunities for interoperability and information exchange.
- Node operation and training: develop guidance, training and an adoption model enabling Node staff to use, understand and maintain their part of the system.
PrincipleTechnical integration should be pursued where it creates clear value and suitable interfaces exist; not every local system needs to become technically integrated with the Navigator.
Related formal tasksWP5T5.1WP5T5.2WP5T5.3
7
Connect AI4Access with INFRA-SERV user-access projects
AimUse AI4Access to improve navigation across Euro-BioImaging access opportunities without creating unnecessary technical dependencies between projects.
- Map the actual access processes of the new INFRA-SERV projects once these are operational.
- Determine what information the Navigator needs to guide users towards appropriate funding/access opportunities.
- Distinguish between requirements for information flow and cases where genuine technical integration would provide additional value.
- Develop and test selected pilot connections later in the project once both sides have sufficiently mature processes.
NoteThis activity can remain lighter during the initial implementation phase.
Related formal tasksWP2T2.5
8
Establish trustworthy and responsible AI operation
AimEnsure that the Navigator can be used operationally by Euro-BioImaging with appropriate safeguards, transparency and staff understanding.
- Clarify application of the EU AI Act and other relevant regulatory requirements.
- Develop practical Euro-BioImaging policies for AI-supported access provision, including human oversight and responsibility for recommendations.
- Define data-safety, privacy and provenance requirements.
- Review implications for the Euro-BioImaging Access Policy and related operational documentation.
- Develop appropriate AI literacy/training for Hub staff and, subsequently, Node staff.
PrincipleTrustworthy-AI requirements should inform the knowledge architecture and validation approach from the beginning, rather than being added after technical development.
Related formal tasksWP6T6.1WP6T6.2WP6T6.4
Bio-Hub's own internal targets — distinct from, and generally earlier/more granular than, the formal deliverable and milestone dates on the tracker.
M1–M3Agree initial target users/use cases; assemble query evidence; start Node engagement; map existing data
~M3–M6Draft minimum information requirements; prototype data structure; define test case; start validation criteria
M6Enough real structured data should exist internally to start meaningful technical testing
M12Minimum data model + PID strategy; first integrated Navigator prototype; Bio-Hub image-analysis landscape
M18Initial Node dataset + engagement toolkit; Bio-Hub facility-system landscape; first KTH AI backend; privacy/trust architecture
M24Integrated Node + website knowledge package; Bio-Hub functional image-data integration module
M30Bio-Hub end-to-end imaging→analysis→deposition use case; transparency/explainability operational
M36Real-user validation + Node deployments + training/adoption model + INFRA-SERV pilots + regulatory/operational readiness