Solution guide · Proposed implementation. The records, examples and acceptance criteria below are illustrative design proposals, not delivered customer systems or measured results.
Replace product persuasion with accountable interpretation
A visitor scans a label because something caught their attention. They may want a short introduction, a deeper explanation, a description of visual details or help finding the next gallery. A useful digital guide should support these different needs without inventing an object's history or forcing everyone into a voice conversation.
AskShelf offers a retail interaction pattern; its setup page documents QR-linked pages and voice entry. A museum implementation would be new software with a different content model, editorial responsibilities and accessibility requirements. This guide is a proposed architecture, not an existing AskShelf museum deployment.
The central change is authority. A museum guide should answer from approved interpretive material and make uncertainty visible. It should not optimize for purchase persuasion, generate provenance from visual resemblance or present a plausible story as a curatorial conclusion.
Create an exhibit record with editorial ownership
| Proposed exhibit field group | Public-content boundary |
|---|---|
| Identity and location | Accession or exhibit ID, public title and display location |
| Interpretation | Approved short/extended material and source notes |
| Language and access | Language versions and accessibility description, each with approval context |
| Rights | Media permissions, credits, expiry and restrictions at asset level |
| Ownership and revision | Editorial owner, source revision, review date and publication state |
Keep internal collection-management information separate from material approved for public display.
An accession identifier and a temporary exhibition label may have different lifetimes. Model their relationship explicitly so a relocated object does not retain incorrect gallery directions. QR destinations should resolve through a managed mapping that can show a relocation notice without changing the object's scholarly record.
Every language version needs its own approval status and relationship to the source revision. When curators revise the base interpretation, mark dependent translations for review. Do not assume that a previously approved translation remains current after a date, attribution or qualification changes.
Rights belong at the asset level. An approved text does not automatically authorize reproducing every associated image, recording or archival document. Record allowed uses, credit requirements and any expiry or restriction, then prevent unsupported media from appearing in the guide.
Offer multiple routes to the same interpretation
A proposed visitor flow starts with a QR scan or printed short code. The landing page immediately offers a concise text introduction, language selection and optional audio. Visitors can explore approved detail or ask a question; the response should identify its supporting material or say the collection record does not answer it.
Keep the basic interpretation available without a conversation. A visitor should not have to grant microphone access, create an account or phrase the right question to read the label's accessible equivalent. Provide a route from the museum's main website for people who cannot scan the code.
Include accessibility specialists and disabled visitors in the design and evaluation rather than treating a checklist as a substitute for lived experience. This is a proposed evaluation process, not certification of the guide.
Digital controls should be designed and tested against applicable WCAG 2.2 criteria, including appropriate text alternatives, keyboard access and accessible names. WCAG concerns web content and does not, by itself, establish accessibility of the whole museum visit.
An illustrative disputed-attribution failure
Imagine a fictional exhibit called Painted Vessel E-17. Its approved record says the workshop attribution is uncertain, while an older archived label gives a definite maker. A visitor asks who made it. A retrieval system finds the older text because it contains a clear name and returns that as fact.
The proposed content service filters by public approval and current revision before retrieval. It returns the uncertainty expressed in the current record and links to the approved explanation. The older label remains available to authorized editors for historical context but cannot override the published interpretation.
Now imagine that one translation still reflects the withdrawn attribution. The guide should not silently serve it as current. It can offer a reviewed short alternative, the current source-language text or a clear notice that the translation is being updated. Curatorial uncertainty must survive translation, summary and audio generation.
This is an illustrative editorial failure scenario, not a claim about any institution's collection. It demonstrates why source ranking alone cannot replace publication state and editorial ownership.
Make descriptions and audio deliberately different
An audio reading of a label is not necessarily an adequate description of an object. Curators and accessibility reviewers should decide which visual details need description and how those details relate to the interpretation. Keep descriptive content identifiable rather than mixing observation and historical inference in one generated paragraph.
Offer playback, pause and replay controls that work with assistive technology. Provide text equivalents for the information in audio, and avoid automatic playback that competes with the gallery environment. Captions and other alternatives may be needed for audiovisual material, depending on its content.
Language choice should persist as visitors move between objects, but the system should state when a particular object lacks an approved version. Never fill that gap by presenting an unchecked translation as curator-approved. Names, contested terminology and culturally specific concepts deserve human review.
Plan for connectivity and visitor privacy
A low-bandwidth first screen should deliver the essential approved text before optional large media. Where offline caching is implemented, record the content revision and make outdated material replaceable on reconnection. Avoid promising that every interactive question will work offline when the answer service requires a network connection.
Visitor services information needs separate operational ownership. Opening times, temporary closures and directions should come from current museum records. An exhibit interpretation source cannot establish whether a lift is working or a gallery is currently accessible.
Collect only the usage information needed for a defined purpose. Aggregate page use can help identify broken routes without keeping identifiable conversations by default. Children's visits require particular care: avoid unnecessary identity collection, targeted profiling or linking questions to named visitors. Review applicable privacy obligations locally rather than treating this design proposal as legal compliance advice.
Evaluate interpretation, access and recovery
Build an editorial test set with answerable questions, genuinely unknown provenance, withdrawn labels, conflicting revisions and requests beyond the curated material. Curators should mark which claims are supported and which qualifications must remain. Measure unsupported assertions and lost qualifications separately from visitor satisfaction.
Run voluntary usability sessions with visitors using different languages, access methods and assistive technologies. Test short-code entry, keyboard navigation, screen-reader output, audio controls and recovery after poor connectivity. Ask whether visitors can reach the same essential information without speaking or scanning.
Pilot a small, well-documented group of exhibits with named editorial owners and a rapid withdrawal process. No interpretation or accessibility improvement should be claimed as measured until the relevant review or study has actually occurred. A clear unknown response is a successful outcome when the collection evidence is incomplete.
For a scoped project, review Datahase's web-development services and share a visitor-guide brief. Bring approved sample labels, rights information, language priorities and access needs. Those materials determine a responsible first release more directly than the number of exhibits alone.
Product foundation and implementation boundary
The AskShelf owned-product study follows its documented catalog, QR, mobile-page and browser-voice workflow. That retail pattern informs these proposals; it does not establish the domain-specific records, authoritative matching, approvals or integrations described here. Those require additional engineering and operational ownership. Public documentation is not an independent product test, a performance guarantee or evidence of a delivered sector system.
Sources and applicability
Public sources checked 5 November 2026: the linked AskShelf workflow and feature documentation, plus WCAG 2.2, used for web-content alternatives, keyboard access and component names. It does not establish accessibility of a whole museum visit or certify this proposed guide. Implementation suitability and proposed release gates require their own review and testing.