A retrieval-augmented assistant embedded inside every Accadema product. EU-hosted, no training on customer data, always cites the source.
Athena is the platform's AI surface. It is embedded inside Atlas for discovery, inside Stellaris for research-office work, inside Tesara for repository navigation, inside Apolon for analytics narration. Every answer is retrieval-augmented over the institution's own data — the institution's catalog, the repository, the CRIS — and every answer carries a citation back to the underlying record.
Customer data is never used for model training. The retrieval index belongs to the institution; access is governed by the same authentication service as the rest of the platform. The model itself runs in EU infrastructure, in line with the platform's data-residency posture.
One assistant, every surface. Catalog Q&A in Atlas, research summaries in Stellaris, deposit navigation in Tesara, narrative analytics in Apolon.
Retrieval-augmented over the institution's own data, with citations attached to every answer. No hallucinations smuggled in as confidence.
Inference in EU infrastructure. No training on customer data. Permissions inherited from the same identity service as the rest of the platform.
Athena is the AI surface that does not separate the answer from its source — and does not separate your data from your institution.