Image source: Robert Scoble.

The mobile Internet experience is vastly different from accessing the web on a desktop computer. People use apps instead of a browser, and that means the companies that own the most popular apps are at a significant advantage. According to Nielsen, Facebook (META 2.98%) and Google own all eight of the most popular apps in the United States. But Facebook likely dominates total engagement compared to the Alphabet (GOOG 1.25%) (GOOGL 1.27%) company. Facebook management says its apps account for 20% of time spent on mobile.

Now, Facebook is encroaching on Google's territory, offering improved search functionality in its flagship app, and experimenting with search in Messenger, which has 700 million global users. Google is reportedly responding to Facebook's efforts with a messaging app of its own featuring a chat-powered search engine.

Can Google stop Facebook?
Google's planned mobile messaging app sounds like a direct response to Facebook's M. M is an AI-powered virtual assistant that can do anything from answer simple search queries to book travel accommodations. It's still limited to select users in the Bay area, but it represents yet another example of using Messenger as a platform.

As M rolls out to more users, Google could find itself losing out on search queries. The planned messaging app is aimed to prevent such an occurrence. But Google is missing out on the step that enabled Facebook to threaten Google's cash cow in the first place.

Google needs to build a network. People use a messaging app because all of their friends already use it. The network effect is quite apparent in the regional dominance of chat apps like WhatsApp, Messenger, WeChat, Line, and KakaoTalk.

Building a messaging service around several AI-powered chatbots is unlikely to draw the critical mass necessary to make a messaging app succeed. Google has failed to properly grow social networks in the past (see Google+), and its existing messaging apps, Hangouts and Messenger, have failed to attract a significant audience. Adding chatbots that answer simple searches won't change that.

What Google could do instead
Google is certainly capable of rivaling Facebook when it comes to artificial intelligence. Last year, Google hired Geoffrey Hinton, one of the world's foremost researchers of deep learning neural networks, to help it design new AI algorithms. The move came just after Facebook hired NYU professor Yann LeCun to head up its AI research lab.

But, as mentioned, Google is missing the network. So, why not launch a chatbot on Facebook's platform? Facebook opened Messenger to third-party developers in March, enabling users to insert pictures, videos, and other media from other apps. Google could develop its AI-powered chatbot for Messenger.

Google has had no qualms developing software for platforms owned by competitors. It has dozens of apps for iOS, for example. But developing for Facebook's platform might not be in Google's best interest, as it cedes control and provides data to a company that uses the same monetization model as itself. Thus, Google is left to its own devices to create an app that people will use -- and use to search.

Mobile search growth
Earlier this year, Google announced it receives more searches on smartphones than it does on desktops. To be sure, the number of queries on mobile devices continues to rise, and those queries are largely Google's to lose. Facebook's M and its flagship app's search capabilities represent real threats, but development remains slow and limited. Google has time to figure out how to compete with Facebook to maintain its share of the mobile search market as it continues to grow.