M's logo. Source: Facebook.

The recently announced M virtual assistant from Facebook (META 2.98%) has gotten a lot of press since David Marcus, VP of messaging products, unveiled it in a Facebook post. Most of the articles focus on how it compares with services such as Apple's Siri, Microsoft's Cortana, and Google's (GOOG 1.25%) (GOOGL 1.27%) Google Now, all of which provide the same basic digital voice assistants.

These assistants are most useful when you need to interact with your phone -- check your calendar, set a reminder, send an email. They're also capable of performing searches with a limited scope -- sports scores, weather, math problems.

M is a whole other beast. It allows users to search with unlimited scope and interact with everything outside their phone. What's more, the service is personalized to you. This difference makes it a much bigger threat to standard search engines such as Google and Yelp (YELP 1.57%) than it does to Siri and Cortana.

What can M do for you?
I'm a big fan of Siri. I use it to set timers, queue up music, and send texts, but when I want to find some information, it's pretty bad. I more often find myself navigating to Google to find answers to questions or receive suggestions.

Source: Facebook.

It's that secondary function, the one I typically use Google for, that M wants to take over. It uses a hybrid of artificial intelligence and human "trainers" to produce restaurant recommendations, help you shop for gifts, find airline tickets, and a host of other tasks. With the help of humans, it can book reservations at those restaurants, order those gifts, and book those tickets.

What's more, M learns about you the more you interact with it, so its suggestions become more tailored toward you and your personality. Facebook says it's only using data gathered from conversations with M at this time, but that may change in the future.

If Facebook does update its privacy policy to allow M access to all of its personality data, it could have some of the most personalized suggestions for users. A recent study found that it's possible to create a personality model based on Facebook likes that's more accurate than a personality assessment from friends and family. Only a few other companies have enough data to provide similar personality profiles, Google being one of them.

Google does use that data to provide somewhat more personalized results, but it typically uses it to target advertisements to users in search results and around the Web. With the curated single-result approach of Facebook's M, personalized results could be a major difference-maker as to whether the service is useful or not.

Source: Facebook

A limited scope, but the most valuable
M's use is still limited, but it's limited to some of the most valuable verticals. Yelp built its entire business on local restaurant searches and recommendations. It added the ability to make reservations in 2010 and brought that in-house with the acquisition of SeatMe in 2013. If M can do a better job of recommending restaurants using Facebook's data, then Yelp is going to see a drop in searches and reservations.

Retail is another vertical where M can offer quite a bit of assistance. Retail websites contribute a huge portion of Google's ad revenue, and that number is only continuing to climb with the popularity of Google's shopping ads, originally introduced in 2012 as product listing ads. Last year, the company saw an estimated 47% growth in ad spend for shopping ads in the fourth quarter. Retailers are by far the biggest digital ad spenders in the United States.

The last major vertical M could take a serious chunk out of is travel. The travel industry doesn't advertise quite as much as retailers; its preferred method is through direct-response ads, which is a market Google's search ads dominate. RBC analyst Mark Mahaney estimated that Priceline and Expedia alone would account for about 5% of Google's total ad revenue in 2014.

If M finds success in any one of these markets, it's bad news for Google or Yelp. Then it's up to Facebook to figure out how to make money on it.