The National Gallery Prague cares for the most significant art collection in the Czech Republic and aims for open, inclusive communication with its audiences. Our goal was to design the audioguide as a system, not a one-off app: a web-based PWA platform capable of handling multiple languages, voices, and – soon – different levels of narration (children – general – expert), while maintaining curatorial quality. We call it a dynamic context.

• Installation-free experience, seamless use on-site within NGP.

• Balancing curatorial quality with scalability.

• Diverse audiences and situations.

• Maximum simplicity in user experience.

• Web-based PWA, no installation required.
The audioguide runs directly in the browser, so starting the experience takes seconds – scan a QR code, select an exhibition, and start listening.

• AI for authenticity and efficiency – not for effect.
We use AI for text-to-speech, multilingual output, and stylistic variations, enabling curatorial content to be quickly adapted to different audiences. Some audio tracks remain recorded by actors to preserve artistic nuance.

• Dynamic context.
One content base, multiple perspectives and levels of complexity, allowing visitors to experience diverse viewpoints and develop critical thinking through interpretation.

• Curatorial interface.
An internal dramaturgical tool – curators and editors retain meaning and quality while AI supports adaptation.

• UX designed for approval workflow.
An admin interface tailored for managing AI-generated content through review and approval steps.

• AI × AQ in practice.
AI describes – AQ helps understand. In our system, they collaborate: AI scales the content, curators preserve interpretation and meaning.

• Multiple voices and languages at a click.
Synthetic voices make it possible to adjust tone and tempo to the visitor’s profile without complex dubbing workflows.

• Modular architecture.
From wireframes and audio-stop widgets to CMS and curatorial tools – exhibitions and languages can be added easily.

• Higher adoption than native apps.
The PWA audioguide targets 10–15% of visitors, compared to 2–5% typical for traditional solutions; personalization further boosts engagement.

• Longer “time-with-art.”
Visitors using the audioguide spend 2–3× more time in the exhibition – more attention means greater value of the visit, and even higher conversion in the gallery shop.

• Scalable content production.
AI reduces preparation time for multilingual tracks and alternative versions, allowing the team to respond faster to new exhibitions and easily update existing content.

Together with the National Gallery Prague, we designed and developed the concept, UX/UI, curatorial workflow, and AI module (text → voice → language). Our shared goal: to build a sustainable system that strengthens the institution’s mission – to open art to the widest possible audience and foster critical understanding.