SEATTLE — A recent hands-on experiment with Google's AI-powered assistant, Gemini, integrated into Google Maps has shown promising results for users looking to plan personalized day trips in urban environments. Allison Johnson, a technology writer for The Verge, tested the feature by allowing it to curate a full-day itinerary around Seattle, relying on public transit and focusing on lunch, a scenic walk, and a laptop-friendly coffee shop. The trial, conducted on a rainy day in the city, highlighted both the tool's strengths in surfacing hidden gems and its occasional pitfalls, such as minor inaccuracies in directions.
Gemini, Google's large language model, has been rolling out across various company services over the past year, often appearing in tools like Gmail whether users opt in or not. Its integration into Google Maps, however, is a newer development aimed at enhancing exploration and planning. Johnson, an avid Google Maps user who frequently scouts new spots like bike routes, playgrounds, and coffee shops via the app, decided to challenge Gemini to venture into unfamiliar Seattle neighborhoods. She specified parameters including public transit travel, visits to two distinct areas, and a return home by 4:30 p.m.
The itinerary began with lunch at Tacos Chukis, a tucked-away spot in a building shared with several other retail shops, lacking a prominent sidewalk sign. According to Johnson's account, Gemini guided her precisely to the location, which had just opened 15 minutes before her arrival. She opted for the house specialty featuring grilled pineapple, describing the three tacos as excellent. This stop underscored Gemini's ability to draw from user reviews and other data sources within Google Maps to recommend popular items.
Ahead of schedule after lunch, Johnson prompted Gemini for a nearby unique shop to browse before heading to her next destination. The AI suggested Elliott Bay Books, a well-regarded bookstore, but inaccurately claimed it was just one block east — in reality, it was a 10-minute walk in the opposite direction. Johnson noted this as the only significant hallucination during her experiment, especially frustrating amid pouring rain. After correcting the AI, it redirected her to Kobo, a store specializing in Japanese goods located nearby. Johnson, who had visited Kobo's other location before, appreciated discovering this outpost.
From there, Johnson proceeded to Volunteer Park for a nice walk, arriving with a soaked jacket despite her umbrella. Gemini offered options for a scenic loop around the park or a visit to the Volunteer Park Conservatory, a large greenhouse. Opting for the latter to escape the weather, she paid a $6 admission fee — a detail Gemini had omitted, which mildly irritated her. Inside, Johnson marveled at the tropical plants, including a tree that hollows itself out to house protective ants.
A conservatory staff member noticed Johnson's enthusiasm while she photographed towering palms and led her to the cacti room. There, the worker insisted on taking Johnson's photo amid the giant cacti and remarked, “It’s so beautiful in here I could cry!” Johnson agreed, describing the cacti as haunting and mystical, with varieties ranging from fluffy specimens to those resembling peeling wax paper or cartoonish saguaros. This unexpected indoor oasis transformed a dreary Pacific Northwest day into a memorable experience.
Reflecting on her visit, Johnson purchased small mementos from the gift shop: a soggy admission ticket and kid-sized gardening tools. She connected this to her parenting life, noting how her child enjoys digging in the yard, and pondered simpler joys like playing in the dirt at home rather than always seeking out new excursions. This stop emphasized Gemini's role in facilitating serendipitous discoveries, even if not all details, like the entry fee, were perfectly conveyed.
The day's final leg took Johnson via the Route 10 bus to Day Made Kaffe, a Scandinavian-inspired coffee shop in a neighborhood she frequents but had somehow overlooked. Gemini described it as minimalist yet warm and laptop-friendly, which aligned with Johnson's preferences. Upon entering, she realized the space was formerly a fancy home goods store where she had shopped for Christmas gifts in 2024. The AI recommended a cardamom bun, but it was unavailable, so she chose a guava jelly pastry instead.
At Day Made, Johnson enjoyed strong coffee and impeccable vibes, watching the Artemis II launch on mute while working. She departed at 3:40 p.m. as planned and caught her last bus, arriving home at 4:26 p.m. — just four minutes ahead of her deadline. Johnson credited the success not to Gemini alone, but to the human-generated reviews and recommendations that underpin Google Maps' vast dataset.
Gemini's functionality in Maps appears as “Ask Maps,” featuring a text box for queries. It leverages Maps data, including user reviews, and can incorporate external information like weather forecasts. For instance, if asked about umbrella needs for a trip, it checks current conditions. Johnson appreciated how the AI shows its reasoning process for suggestions, building some transparency. However, she cautioned about hallucinations, particularly in real-world navigation where errors could lead to inconvenience, as seen with the bookstore mix-up.
This experiment comes amid broader discussions about AI's role in daily life. Johnson referenced a conversation with her friend Will Sattelberg from 9to5Google, noting how tech demos often push toward transactions like booking flights or buying sneakers. She examined her own habits of seeking purchases — books, coffee, trinkets — as a way to combat anxiety over endless urban choices. Yet, Gemini proved useful for filtering overwhelming options, such as finding kid-friendly spots with specific amenities like high chairs and craft cocktails.
Importantly, Gemini does not generate freestyle transit directions; it integrates with Google Maps' accurate, real-time routing. This hybrid approach ensures reliability for getting from point A to B, while the AI handles discovery and planning. Johnson's trial focused on leisurely exploration but highlighted potential for practical scenarios, like quickly locating a nearby restaurant during a family meltdown.
The integration of Gemini into Google Maps reflects Google's push to embed AI across its ecosystem, following its debut in services like Gmail over the past year. While some users may find the omnipresence intrusive, early tests like Johnson's suggest value in personalized, data-driven recommendations. As AI tools evolve, concerns about accuracy remain, but features showing sources and integrating verified data could mitigate risks.
Looking ahead, Google has not detailed further expansions for Gemini in Maps, but ongoing updates could refine its hallucination issues and enhance cross-service integrations. For urban explorers like Johnson, who dreams of endless wandering to bookstores and stationery shops, such tools offer a way to combat decision paralysis without replacing the joy of human-curated discoveries. Her day out, blending tacos, plants, and coffee, demonstrated how AI can serve as a helpful middleman in navigating the city's endless possibilities.
In Seattle's rainy climate, where public transit and spontaneous plans often intersect, innovations like this could encourage more residents to venture beyond familiar neighborhoods. Johnson's experience, captured through photography she took during the outing, provides a relatable snapshot of AI's real-world application, balancing excitement with practical caveats.