How do the Google Assistant’s new features compare to Amazon Alexa’s?

Google announced a bunch of updates for its Assistant yesterday at I/O, including new ways to interact with Assistant and new voices. The announcements will likely improve the experience of using Assistant, but how do the features stack up against Amazon’s Alexa? We break down each of Google’s Assistant-related announcements from yesterday and how they compare to what Alexa can already do.

Google says that it’ll support more than 30 languages and will be in 80 countries by the end of the year. It also says Assistant is now compatible with more than 5,000 home devices, which is up from only 1,500 in January, and is installed on more than 500 million devices. It’s also in cars from over 40 auto brands.

Amazon says more than 12,000 devices work with Alexa, and it’s available in BMWs, select Mini vehicles, and Ford cars. Google obviously has an advantage in the car space because of its successful Android Auto program. Amazon has mainly pushed its Alexa assistant to users through the company’s own hardware, like the Echo and Echo Dot, but it has also partnered with third-party makers. Alexa only supports three spoken languages at this time: English, Japanese, and German.

Users have some flexibility when it comes to issuing commands to the Google Assistant. With the new continued conversation, users can keep asking questions without having to say “Hey Google,” every time. The new multiple actions feature lets them ask multiple questions at the same time. Both of these should allow people to talk to their assistants more naturally, although Google still only allows users to say “Hey Google” or “OK Google” to wake up Assistant.

Amazon recently launched its follow-up mode, which is essentially the same as Google’s continued conversations. You can ask multiple questions, so long as you do so within five seconds of the initial wake-up command. This follow-up mode is opt-in and is available for all devices in the Echo lineup, as well as some third-party devices. It’s also only available in English at the moment and only works when Alexa isn’t being used for another persistent activity, like listening to music or an audiobook. Alexa doesn’t really have an answer to Google’s multiple actions, apart from multistep routines, in which a specific phrase, like “good morning” can trigger multiple actions. Users can change Alexa’s wake word to Echo, Amazon, or Computer on Amazon’s own Echo speakers, but not on third-party devices.

Google announced “pretty please” yesterday, which encourages kids to politely ask the Assistant for things by saying “please” and “thank you.” The Assistant will thank kids for saying please and will call out when they have asked for something nicely. Google launched family experiences last year, which includes games, activities, and stories for families with kids. Parents can also make kids under 13 their own Google account so that they can monitor their usage and set time limits.

Amazon just launched its Echo Dot Kids Edition that comes preloaded with the company’s FreeTime, which is essentially a bunch of white-listed content that’s child-appropriate. The basic FreeTime on Alexa service blocks certain Alexa features, like shopping, news, and third-party skills that require linking external accounts. It also allows parents to set time limits on when and how much Alexa can be used in a day and to pause access to Alexa for periods of time. As part of that FreeTime Alexa service, the company is including a feature called Magic Word, which provides positive reinforcement whenever kids use the word “please.”

Google is rolling six new voices out to the Assistant, including one based on singer John Legend. These new voices join the generic female and male-sounding voices that users could originally use. These six new voices are now available on Android devices through the Assistant settings. Legend’s voice should be coming soon.

Alexa has multiple voice options, including female and male variants with an American, Canadian, Australian, Indian, or English (UK) accent. It also has different intonations and can whisper, vary its speaking speed, and bleep out words.

Google is planning to revamp the look of its Assistant’s smartphone app to make it more like Google Now. Users will be able to view a full page of information related to their query. On smart displays, Google Assistant can play content from YouTube, which Echo devices, like the Show, cannot. The company is also bringing its Assistant to Google Maps to help with navigation. Alexa can provide traffic information but it won’t give turn-by-turn directions in the car like Google Maps or Waze.

One of Google’s biggest announcements yesterday was its work on Duplex, which will let the Assistant make phone calls for users and schedule appointments. The idea is that busy adults might need a doctor appointment made or a salon appointment booked, so the Assistant can handle a conversation with humans to find a date and time. The company hasn’t announced any sort of timeline for its release, however, so we don’t know when it’ll actually come to users.

Amazon’s Alexa does nothing even close to this, although users can buy an Echo Connect to make phone calls from their Echo device as if it were their home phone. Alexa users can also call most phone numbers from their Alexa-equipped speakers. They can also call other Alexa users through tablets or the Alexa app on their phone.

Generally, it seems like Google is trying to make its Assistant as conversational and useful to users as possible. The Duplex feature, while a little unsettling, could fundamentally change how we interact with and use our smart assistants. Although Amazon has a head start on smart home device compatibility, it might not have the same talent or resources to pull off these true AI assistant features and will instead need to rely on independent developers to do the heavy lifting.

Update 5/9, 4:11 PM ET: Updated to reflect that the Google Assistant’s new voice profiles are now available.