Decision support that explains results and flags uncertain cases.
Your staff need to know why a model produces a result, what alternatives exist and when further review is needed. A number alone does not answer those questions.
IAP combines the analytical method with rules and visualisations that show the features and relationships behind a result. Together with your experts, we define when a recommendation can be used and when an uncertain case should be referred for review.
You receive an application that presents results alongside explanations and indications of uncertainty. Your staff can investigate cases and make the decisions or approvals assigned to them.
Read the technical detail ↓A decision support system connects data, models and domain rules to a specific workflow. It presents relevant results, alternatives and uncertainty so experts can justify a decision. Its economic value depends on the benefit of that decision relative to development and operating costs.
Explicit decision rules, data structures and case-level assessments provide different forms of explanation. FlowXAI combines such elements in a research application. Explaining a result and assessing its reliability are distinct tasks.
Unsuitable or insufficient inputs must be recognisable. Depending on the task, the system can withhold outputs or flag a case for review. Implementation defines responsibilities and the boundary between recommendation, approval and automated action.
Thrun, M. C., Hoffmann, J., Krause, S. W., Krawitz, P., Stier, Q., Neubauer, A., Brendel, C. & Ultsch, A. (2026). Self-explaining artificial intelligence for the classification of B cell non-Hodgkin lymphoma: A diagnostic decision support study. PLOS Medicine, 23(7), e1004889.
FlowXAI combines assessment of data structures with explainable lymphoma classification and an estimate of reliability for each individual result. This helps domain experts understand results and focus their review on cases requiring closer examination.
Thrun, M. C. (2022). Identification of Explainable Structures in Data with a Human-in-the-Loop. KI – Künstliche Intelligenz (German Journal of Artificial Intelligence), 36, 297–301. Springer.
A methodological foundation for explainable data structures and the involvement of human expertise in the analytical process.
Commissioned development · decision support
Market information needed to move beyond data presentation and provide a basis for daily trading decisions.
Project details →
Commissioned prototype · related collaborative research
Interpreting flow-cytometry data requires expertise. Rare entities, different measurement conditions and limited training data complicate automated classification.
Project details →
Commissioned development · delivered solution
Train operating data was to provide indications of conditions relevant to maintenance. Changes caused by varying loads and operating conditions needed to be distinguished from possible component faults.
Project details →
Proof of Concept (PoC) using real data
Initial review of a property listing requires information from text, images and other data sources. The aim was to automate recurring preparatory work and combine the findings in an overview for assessment.
Project details →Option 1
IAP investigates the problem, develops and evaluates the approach, and hands over the agreed prototype with documentation. Your IT team takes responsibility for further product development and operation.
Option 2
IAP develops the agreed application, including its interface and integrations, connects it to the intended workflow, and completes acceptance and handover. Your organisation or operating partner takes responsibility for subsequent operation.
In a free, non-binding 20-minute introductory discussion, we clarify your objective, starting point and a useful next step.
