I let Gemini in Google Maps plan my day and it went surprisingly well
Foto: The Verge AI
Just six dollars and a brief conversation with a chatbot were enough to rediscover a city while avoiding tourist traps. Gemini in Google Maps, tested as a personal guide, proved capable of going beyond simple route planning by offering precisely personalized daily itineraries. The "Ask Maps" system analyzes not only locations but also thousands of user reviews and external data, such as weather forecasts, to tailor suggestions to specific needs—ranging from laptop-friendly cafes to niche shops featuring Japanese crafts. Although the technology still experiences minor stumbles—such as incorrectly estimating the distance to a bookstore or omitting information about an entrance fee for one of the attractions—its effectiveness in finding "hidden gems" is impressive. Gemini can suggest specific menu items (like pineapple tacos) and react flexibly to schedule changes caused by sudden rain, for example. For users, this means an end to decision paralysis when browsing hundreds of pins on a map. Instead of wasting time on manual research, we can hand over the reins to artificial intelligence, which is increasingly accurate in predicting our tastes. AI in navigation is ceasing to be just a tool for avoiding traffic jams, becoming instead a curator of our daily experiences.
Dialogue instead of a search engine
The key to the new functionality is the "Ask Maps" button. After pressing it, the user sees a familiar chat window. This is where the magic happens: Gemini is not limited to displaying a list of results, but analyzes data from user reviews, opening hours, and even external information, such as the rain forecast. In the test, specific conditions were set: travel by public transport, a stop for lunch, a walk, and a cafe suitable for working with a laptop – all with a strict requirement to return home at 4:30 PM.
Initial suggestions were quite conservative, but after a short exchange of sentences, the system proposed a route including tacos, a visit to a conservatory, and a Scandinavian cafe. Importantly, Gemini demonstrated great time precision. In the case of the Tacos Chukis restaurant, the assistant directed the user to the right place (despite the lack of a sign on the street) exactly 15 minutes after the premises opened, which allowed her to avoid queues. The system also correctly suggested the house specialty – tacos with grilled pineapple, based on popular opinions from other guests.
Hallucination traps and digital instinct
Despite its advancement, this technology still struggles with the problem of so-called hallucinations. During the test, Gemini claimed with full conviction that the famous Elliott Bay Books bookstore was located just one block east of the current location, which was an obvious error. Had it not been for the user's alertness, the AI would have sent her on a 10-minute walk in pouring rain in the completely opposite direction. After having the error pointed out, the system quickly corrected it and proposed an alternative in the form of the Kobo shop with Japanese merchandise, which turned out to be a bullseye.
Read also
This incident sheds light on the current limitations of LLM (Large Language Models) in navigation. Although they can synthesize huge amounts of text data, they can still make mistakes in the topography of the terrain. Therefore, Google uses a safety switch: when it comes to actually moving from point A to point B, Gemini does not generate its own directions, but launches the standard Google Maps navigation module with real-time data.
AI as a curator of experiences
The most interesting aspect of the test was the visit to the Volunteer Park Conservatory. Gemini suggested this place as an escape from the rain. Although the assistant omitted the fact that entry costs 6 dollars, the recommendation proved invaluable. The visit to the glass conservatory full of exotic plants and cacti became the highlight of the day – an experience that the user might not have sought out on her own, focusing on typical consumer goals.
- Personalization: Gemini can search for places with a very specific profile, e.g., child-friendly restaurants with a vehicle theme.
- Review synthesis: Instead of reading hundreds of reviews, the user receives ready-made conclusions regarding the most popular dishes or the atmosphere of the place.
- Logistics: The system can plan a return to the minute – in the case described, the return home took place at 4:26 PM, just 4 minutes before the set deadline.
The final stop at Day Made Kaffe confirmed the high effectiveness of the algorithm in matching the aesthetics of the place to the user's taste. The minimalist decor and an atmosphere conducive to working with a laptop were exactly what was expected. Interestingly, the system suggested a specific pastry – a cardamom bun – and when it was sold out, the user realized that this place was located on the site of a former interior design shop that she had visited in 2024. Gemini effectively connected the dots in the urban fabric that a human might not have noticed.
The implementation of Gemini in Google Maps is more than just another chatbot feature. It is an attempt to tame information chaos. Google Maps has a gigantic collection of data that becomes overwhelming for the average user. AI plays the role of an efficient intermediary here, doing the "dirty work" for us by filtering thousands of reviews and opening hours. Although we still have to remain vigilant against topographical mistakes, this technology is realistically changing the way we explore our surroundings, turning Maps from a navigation tool into an intelligent city guide. The effectiveness with which the system can match a daily schedule to rigid time frames suggests that soon planning trips without AI support may seem archaic to us.
