I tried talking to Google Maps with Gemini, and it actually understood what I wanted

(Image credit: Adobe Stock)

I started with a walk to the nearby park. Along the way, I tapped the Gemini button where the Assistant button once dwelt, and asked the AI to “find the best coffee shops along the way, and tell me which one has the highest review score.”

So much for nearby things, but what about longer-term planning while on the road? I set up a drive from my place to an art gallery about an hour away, then started driving.

Sign up for breaking news, reviews, opinion, top tech deals, and more.

As I drove, I asked Gemini in Maps to “suggest the top three scenic lunch spots along the way to my destination that aren’t too expensive and are easy to park at.” Normally, this would mean either opening Gemini or doing more traditional keyword searches in Maps.

Instead, Gemini started describing three places along the route, with a short explanation of why it qualified as scenic, the kind of food there, and threw in their star rating as well. None required meaningful detours. Gemini seemed to understand that scenic only mattered after considering the route and even what scenic meant, with mentions of river views and popular terraces. It even offered suggestions like “parking is easiest on weekdays.”

Traffic updates are something Maps has always handled reasonably well, but I wanted to see how well Gemini’s real‑time awareness could handle complex conditions and adjust plans.

You may like

While on the road, I deliberately took a slower route to observe the response when I asked Gemini, “Tell me if there are any traffic slowdowns ahead, and suggest alternative faster routes that also pass by a gas station or rest stop.” Gemini described an upcoming slowdown on the highway due to roadwork, which has been standard, but then it went well beyond that tradition by proposing two alternative routes.

The options were not only faster under current conditions but also met the additional needs I mentioned. One went past a conveniently located gas station right off an exit, and the other routed through a small town with a rest area that had a snack stand. What stood out was how Maps layered multiple criteria into a coherent set of choices. It balanced convenience and speed by outlining estimated times and clearly marking amenities on the map.

Gemini has been thoroughly embedded in Maps well beyond just updating the Assistant. It’s more like using Gemini with a map add-on than Maps with an AI addition. The casual language made using it feel easier than the more traditionally stilted way of engaging with Maps, and the AI could remember what I had been saying earlier during the drive, so that I didn’t even have to repeat myself about what I was looking for from requests for nearby lookouts or rest stops. It’s more conversational, but also more decisive. It narrows options instead of overwhelming you.

There were moments when results took around 15 extra seconds to load, but nothing that derailed the travel. I didn’t mind it when the results were as useful as they were. Gemini is advanced enough that, when combined with Maps’ data, it seems to understand the spirit of what you want and fills in the rest with context. It also showed some restraint in surfacing suggestions when they are relevant and staying quiet otherwise.

The upgrade changed how I talk to Maps, giving it greater confidence when handling multiple requests. That’s a real change, and one that people will likely be keen to adopt. And, of course, encouraging people to travel toward Gemini in other contexts as well is a destination Google will be happy to provide navigation for.