Celebrating 15!

Museum Navigation with AI

Experiments on how AI could help visitors find art they love.

The Met MapThe Metropolitan Museum of Art

Wayfinding is a Challenge

The Met's collection is massive, spanning 5,000 years of global culture. The physical museum is similarly huge, and its scale can be a double-edged sword. It's easy to get overwhelmed with choices, especially if you don't arrive with a plan.

Navigation with GeminiThe Metropolitan Museum of Art

Our Visitors use AI for Navigation

We've observed an increase in visitors using AI to navigate the museum. This indicates a demand for personalized, real-time wayfinding.

We wondered how we can support them with tools that offer the same flexibility, while also pulling exclusively from our data.

The Met Wayfinding ToolsThe Metropolitan Museum of Art

Logistical Wayfinding

We started by exploring logistical wayfinding, and found that AI struggled with things like mapping routes.

Because there already is long history of traditional algorithmic approaches for these types of issues, we focused on personalized wayfinding instead.

The Met Info DeskThe Metropolitan Museum of Art

"What should I see?"

Our front-desk volunteers told us that people often don't know what the museum offers, and what they actually want to see.

The Met Info DeskThe Metropolitan Museum of Art

Personalized Wayfinding

We wondered how we can help visitors identify areas in the museum that align with their interests.

We collaborated across departments to rapidly develop ideas, and user-test quick prototypes right in our galleries.

Navigation Sprint ProtoypesThe Metropolitan Museum of Art

Prototype Overview

In this article we will look at the following two prototypes:

1. The Heatmap
2. The Compromise Finder

The Heatmap PrototypeThe Metropolitan Museum of Art

The Heatmap Prototype

Our first prototype turns our museum map into a custom heatmap that highlights areas particularly relevant to a visitors' interest.

The Heatmap PrototypeThe Metropolitan Museum of Art

Visitors can search for anything, 'golden jewelry' for example, and explore matching artworks in each room.

The video below shows a visitor searching for different interests. Within seconds, the heatmap overview of their interest is generated. Visitors can browse the entire museum, and explore individual artworks in each room.

The Heatmap PrototypeThe Metropolitan Museum of Art

User-testing showed that the heatmap is great for targeted exploration: Repeat visitors or researchers who have a specific interest in mind.

However, for visitors who don’t know where to start, the prototype presented a "blank canvas" problem by offering too much choice. They preferred a simple first step instead.

The Compromise Finder PrototypeThe Metropolitan Museum of Art

The Compromise Finder Prototype

Our next experiment focuses on places to begin, rather than an open-ended choice of galleries and artworks.

The Compromise Finder PrototypeThe Metropolitan Museum of Art

The prototype is made for groups.

First, each person enters their interests from a broad pre-existing list of categories.

The Compromise Finder PrototypeThe Metropolitan Museum of Art

Then, Gemini suggests a shared itinerary for the group by taking everyone's interests into account.

Below is an example: Two visitors entered their must-sees, and received three suggestions based on their shared interests. AI also found highlights from each department to give them an idea of what they might find.

The Compromise Finder PrototypeThe Metropolitan Museum of Art

Visitors told us they enjoyed the collaborative itinerary that takes everyone’s interests into account.

By selecting galleries to begin with, it also framed the museum as a series of manageable, achievable sections.

The Met Guides and MapsThe Metropolitan Museum of Art

Main Learnings

Our prototypes showed that visitors want to explore museums with more agency and personalization. 

AI can unlock deep customization, allowing for real-time recommendations and suggestions that work at a range of scales, from full-museum down to individual artworks.

Credits: All media
The story featured may in some cases have been created by an independent third party and may not always represent the views of the institutions, listed below, who have supplied the content.
Explore more
Related theme
The Met: Prototypes & Play
How the museum is using AI to explore its collection in new ways
View theme

Interested in Performance?

Get updates with your personalized Culture Weekly

You are all set!

Your first Culture Weekly will arrive this week.

Home
Discover
Play
Nearby
Favorites