Waze is deploying five AI-driven features to its navigation app, spanning motorcycle-specific routing, personalized route suggestions, a reduced-audio "Less Chatty" mode, an expanded Conversational Reporting tool, and Gemini-powered voice search, according to Engadget.

Motorcycle mode uses AI to route riders around hazards that don't apply to four-wheeled traffic — potholes, speed bumps, raised crosswalks, shoulder endings, and narrow bridges — while also surfacing shortcuts unavailable to cars, per Google's description of the feature relayed by Engadget. The underlying data pipeline combines Waze's real-time traffic map with a dedicated cohort of human editors who focus specifically on motorcycle-relevant hazard reporting, keeping that layer of the map current in a way pure crowdsourcing or pure model inference wouldn't reliably achieve on its own.

The feature is rolling out to riders in Argentina, Brazil, Colombia, Malaysia, Mexico, Peru, and the Philippines — a market list that tracks closely with regions where two-wheeled vehicles carry a disproportionate share of daily commuting and delivery traffic relative to North America or Western Europe. Waze has not indicated when or whether the rollout will extend further.

The personalized routes feature layers historical trip data on top of Waze's standard traffic-optimized routing, surfacing suggestions based on where a given user has driven before rather than purely on live congestion and ETA. Users retain the ability to disable it, addressing what has become a fairly standard expectation around opt-out control for behaviorally-informed recommendations. Less Chatty mode trims voice guidance down to only what's classified as critical — turn instructions and safety reminders — stripping out the ambient commentary Waze has historically layered on top of bare navigation.

Conversational Reporting, which Waze began testing in October 2024 as announced on Google's blog, now extends beyond its original scope of verbal incident reporting to include map update suggestions. The feature originally let drivers report hazards or traffic conditions by speaking naturally rather than tapping through menus; the expansion effectively turns the same conversational interface into a crowdsourced map-editing channel, letting users flag stale or incorrect map data by voice while driving.

Gemini-powered voice search is the most explicitly generative addition of the five. It lets users issue natural-language queries — Waze's own example being "Find me a gas station nearby with the lowest prices" — and returns a customized results page rather than a simple list of nearby POIs ranked by distance. This is currently limited to Waze's beta community and is rolling out globally within that tier, rather than to the general release channel. The personalization features, Less Chatty mode, and the expanded Conversational Reporting are rolling out globally on both Android and iOS.

Notably, Engadget's reporting does not point to a single consolidated Google or Waze blog post announcing all five features as one coordinated release. The only Google blog URL cited in that reporting is the October 2024 post introducing Conversational Reporting in its original form. Waze's own blog landing page currently surfaces a post titled "New road reporting features coming to Waze," which references Conversational Reporting, alongside unrelated posts such as "Crash history alerts arrive to the Waze map," "Drive with Ghostbusters on Waze," and "Drive like a champion with Allyson Felix" — none of which function as a unified announcement for this batch of AI features.

That gap is worth flagging for anyone tracking how Google is packaging AI rollouts across its product portfolio. Bundling five separate features — some in limited regional beta, one in global general availability, one in beta-only global rollout — into a single narrative works well for press coverage but makes it harder to pin down a canonical, versioned changelog. Enterprises and developers who build on top of Waze's data (via the Waze for Cities program or similar partnerships) will likely need to track individual feature rollouts rather than treat this as one discrete release event.

The motorcycle mode's country list is the more interesting signal from a product-strategy standpoint. Google has generally treated emerging markets with high two-wheeler density as a distinct routing problem for years, and pairing algorithmic hazard detection with dedicated human editors suggests Waze isn't yet willing to fully hand hazard curation to inference alone in a domain where a missed pothole call has real physical stakes. That hybrid model — AI for pattern-matching and route optimization, humans for ground-truth verification on safety-critical data — is a sensible design choice given the cost of false negatives on a road hazard.

Gemini's integration into voice search also fits a broader pattern across Google's consumer apps this year, where the company has been threading Gemini into search-adjacent surfaces rather than launching it as a standalone destination. Waze's implementation, limited for now to its beta cohort, reads as an early-stage test of whether conversational, intent-based queries hold up in a navigation context where users are typically driving and cannot review multiple results carefully. Whether that beta-only status expands to general release, and on what timeline, is not yet specified in available reporting.