Android Auto messaging is useless, and it's all because of one missing feature

Credit where credit’s due, Android Auto is jam-packed with features, and that’s all thanks to Google’s dedicated development push throughout the years. I remember how clunky and threadbare the driving companion was a few years ago. Now, drivers and passengers can accomplish a wide array of tasks without ever having to get their phones out or even touch the car screen, with even more promised this year.

Despite my appreciation for Android Auto, especially as a daily user, I’m a realist. There are facets of the experience that Google can certainly improve, and one of the most glaring problem areas is its messaging system.

Android Auto, now with the help (or hindrance) of Gemini, makes texting contacts frustrating, distracting, and very often infuriating — and it’s all due to its reluctance to adopt the one feature that’ll remedy it.

Android Auto lacks voice messaging support, and it’s infuriating

Google encourages users to control Android Auto with their voices, whether it’s adjusting navigation plans or sending a message to a contact. Given the latter, you’d expect this to include sending voice messages — specifically audio messages recorded by the car or phone’s microphone — but this isn’t the case. Gemini force-transcribes any message sent to contacts, and this leads to all sorts of issues.

You see, Gemini’s dictation efforts are, in a word, dire. The next-generation voice assistant often misunderstands what I ask it to do, and this is exacerbated when I’m dictating a full message. This workflow usually involves multiple time-consuming steps.

First, I invoke Gemini to send a message to a contact. It’ll then ask for that contact, then ask me to verbalize the message. Gemini will then transcribe it and read it back for confirmation. At this point, after several minutes and multiple rounds of back-and-forth trying to correct the many glaring errors in it — punctuation written out as words, or incorrect names and capitalization — I just give up and say, “Send.”

Of course, it does work sometimes, especially when I’m parked and need to take pictures of its performance for this article. But while I’m driving, when I need it to perform at this level, I run into annoyances. This is when it matters. The in-car messaging experience really doesn’t have to be like this, yet it’s the only messaging capability Google offers through Android Auto.

I don’t expect the system to be reliable in all situations, but it genuinely becomes a guessing game about when it’ll transcribe messages correctly. My fiancée has shared similar frustrating experiences when she pilots our car on weekdays, too.

To Gemini’s credit, it gets “I’m driving,” “See you later,” or “Sounds good,” without issue, but I don’t believe I should praise a fish for swimming. Regardless, it struggles with transcribing more complex messages, and the issue only gets worse the longer it is, the more your accent deviates from traditional US or British English, and the greater the number of uncommon proper nouns in your snippet.

The starkest problem is that there’s simply no workaround to avoid this dictation nightmare. I either bend to Android Auto’s will, place a call, or do something truly dangerous — pick up my phone while driving, and send that voice memo manually. I certainly don’t advocate for the latter, and I usually ask my fiancée to contact anyone on the road on my behalf, but it does explain why I see so many drivers in modern cars with their phones in hand. The current Android Auto messaging system is a safety risk.

The fix to so many of Android Auto’s messaging problems

So how does Google fix this issue? It could, and perhaps should, improve Gemini’s dictation accuracy, but it could take years and countless dollars in development to make it truly accurate in all segments of the globe. The simplest solution is therefore perhaps an obvious one.

Instead of incorrectly transcribing my voice to text, why doesn’t Gemini simply package my verbalization as a voice memo? This way, the recipient receives the exact information I deliver, with the intended tone, punctuation, and context, and without any Gemini-related issues.

Recording my voice verbatim will retain the veracity of my message. And if need be, Gemini could offer users the option to send a supporting transcription if the audio quality isn’t great.

Beyond information delivery, there are other clear benefits that direct voice messaging has over transcriptions. The obvious one is cutting down on the back-and-forth between driver and Gemini, which can be hugely distracting and time-consuming.

Talking to a passenger is something I can do without committing oodles of brainpower — it closely resembles typical human interaction. I can still concentrate on lane changes, avoiding obstacles, and road markings when chit-chatting with my fiancée. Sending a voice message should come as naturally as conversing with a passenger. And since Gemini wouldn’t need to verify message content, the process would be completed far swifter and allow me to focus on the road again.

However, when Gemini asks me questions about my message, I’m pressured into responding in a more timely manner, and then forced to pay attention to what it reads back to me. I have to allocate a greater share of my concentration efforts to this process. Again, this can be a safety issue.

There’s also the argument that most major messaging apps in 2026 support audio memos, whether it’s my primary solution, WhatsApp, or the more traditional Google Messages via RCS. It should, therefore, be a firm part of the modern Android Auto experience at this point.

Redundancy is a wonderful thing

Google should invest more resources in improving Gemini’s dictation accuracy in the car, but this only addresses part of the problem with the messaging experience.

I’m not advocating that Google completely replace its current system with voice messaging, but rather complement the existing offering. Different contexts require different solutions, so giving drivers the option to tell Gemini, “send a voice message” or “send a text message” would allow them to contact others as context demands. Not offering a voice message option forces drivers into using Android Auto in a specific way, one that is clearly fraught with issues, as I’ve explained.

In a recent conversation I had with Gemini on a weekend getaway, I asked it why I can’t send my loved ones a voice message. It answered:

Gemini on Android Auto currently converts your voice to text messages but does not support sending recorded audio clips. This feature is still being developed as part of the ongoing transition from Google Assistant.

At least Google seems aware of the shortcomings and is working to bring voice messaging to the platform. I just hope it arrives sooner rather than later.