Ryan Neil

More than 1.5 billion people use Google Maps and Apple Maps every month. It has become common for people to open a map to plan a trip, look for a place to eat, or navigate. Most consumer mapping applications are available for free with the cost being subsidized by other services, but I won't be covering business models in this post. What I’m primarily interested in exploring is how maps are commonly used and how they can be improved.

Common Use Cases

Consumer mapping applications are most commonly used for two fundamental purposes: searching and routing. Both use cases require adding text to an input field and selecting a location. These days, location search is generally a fast and user-friendly experience with the use of autocomplete enabling accurate results to be returned before the user has finished typing. Routing generally involves searching for multiple locations and selecting a route based on travel mode, time, and cost. Both of these use cases can usually be accomplished without browsing or using another application.

When it comes to trip planning, the complexity increases. Many other queries arise: Where do I want to go? How do I get there? What can I do while I’m there?

Existing apps make it easy to explore locations, look up travel routes, and even make bookings in some cases, but trip planning generally requires a combination of web browsing and other applications. In the end, users will often resort to a text file for documenting and sharing plans.

Trip Planning Example

I want to plan a trip to Normandy. What do I do?

• Start by figuring out how to get to Normandy from my origin. What are the nearest cities with convenient access?
• En route to Normandy, I decide to stay in London for a few days. Where should I stay, and what should I do while I’m there?
• My chosen route to Normandy requires taking the Eurostar to Paris and changing trains. Should I find somewhere near the station to eat lunch?
• The last stop in Normandy is about 30 minutes walk to my local accommodation. Will the weather be nice enough to walk?
• Finally, after planning the key points of the trip, I need to share the itinerary with my fellow travelers. Do we need to make adjustments to accommodate everyone?

The above might take a few hours total of research and map-usage, but generally the process will be spread out over days, weeks, or even months. The user is starting with more questions than answers, and the plan unfolds over many iterations of research, exploration, searching, and routing.

What Can We Improve?

There is room for improvement during exploration, location research, planning, and collaboration. Searching and routing in particular are ripe for innovation given recent advancements in large language models, and personalization is where an LLM can really benefit user experience.

A natural language interface might be used in this scenario to kick off planning in a more vague, exploratory way. Instead of jumping straight into searching for specific locations or routes, a user might start by briefly outlining their trip and key considerations. The LLM can then guide the user with helpful prompts and further context to aid in research. Knowledge about past trips and travel preferences would help tailor the suggestions, and better collaboration with travel companions would ensure everyone’s needs are met.

Considerations

Simply replacing search and routing inputs with a chat interface would not improve the experience for most users. Typing a location name into an input with autocomplete is faster and easier than typing a sentence into a chat interface.

Offering a direct interface for search and routing will ensure users can interact with the map quickly while the chat interface coupled with generative UI components can offer an alternative for exploration, research, and planning. Experimentation and iteration are key to unlocking new opportunities.