From concept to operational scenario
“Having a chatbot” is not an institutional objective. A useful use case describes who needs information, which sources are authorised, what outcome is expected and when a person should intervene.
The following examples represent real and recurring needs in these sectors. They are not presented as customer testimonials or as results already achieved by Acervo.
Museum: add context to an object during a visit
A visitor encounters an object and asks about its period, technique or relationship to other works. The service consults approved collection records and curatorial texts, answers in language suited to the audience and identifies its sources.
Measure: guide usage, explored themes and the proportion of answers supported by curatorial sources.
Local authority: explain a published procedure
A resident asks which documents are required for an application. The service searches current regulations and official pages, presents a summary and refers the person to the relevant desk where individual assessment is needed.
Measure: reduction in repetitive enquiries and correct referral rate.
University: search across repositories
A researcher explores a topic spread across theses, articles and special collections. The service assembles relevant passages, preserves authorship and citation and supports deeper research without replacing source reading.
Measure: time to relevant documentation and coverage of cited references.
Archive: discover relationships across fonds
An archivist searches for occurrences of an entity across separate record series. Semantic search complements finding aids and presents results with provenance, dates and access restrictions.
Measure: result precision and new documentary links validated by the team.
The shared pattern
Across all four scenarios, technology serves a verifiable chain: question, authorised retrieval, answer, citation and, where necessary, human referral. That chain, rather than fluent prose, turns AI into a controllable institutional capability.