Will Apple win the war of AI? Will it even survive in the new world? Plenty of people believe that the answer to both the question is no. I believe, this is the wrong question to ask.

Couple of days ago, Marco Arment penned a post where he likened the current situation to the fate of the Blackberry. He is right in his sports analogy, that Google and Facebook are very good at skating to where the puck is going to be. Amazon is the rest of the team, hoping that by covering the rest of the field, it has a better chance to get its hands on the game. Where is Apple? In the last 18 year, Apple was the guy in the corner quickly melting the ice so he could play basketball. And they have the chance to do it once again. How?


I know what you think: Isn’t this the same thing that everyone else is doing? Well, the difference is in how, not in what. I believe that Apple is in the most unique position to bring to the market a service, that could become the future of search on non-desktop devices.

We all know, that the current version of search sucks (open a website, type couple of keywords, get a list of links, open the first one, find the proper information, copy the text, open an app, paste the information, hope it recognizes it, done). It took us a while to find the obvious solution, but we found it: Personal assitants.

Let’s forget for a moment how disappointed we all are with Siri (or any other personal assistant) and let’s talk about how Apple could make Siri their next multi billion dollar business.

First of all, Siri needs to get much faster. Their goal should be under 0.5s for speech to text recognition. Seems like they’re struggling a lot with this one. One of the options is to acquire SoundHound. Seems like they’ve built a decent technology. It comes with a bonus of music recognition that is currently supplied by Shazam. Shazam’s technology is great, but it would be too expensive to acquire them for what it offers and is painfully dependent on the integration with iTunes. Apple doesn’t need their relationship with music studios, they already have them.

Second, they need to move Siri’s backend to the cloud. First of all, they would get control for most of the “supply chain”. If they only offer an SDK to developers, it does not matter how quick Siri is. It needs to talk to a 3rd party app on your phone, which needs to connect to the 3rd party server that needs to process the data and sends back the results through the same route. They can cut a lot of fat by allowing 3rd party developers to connect directly to their servers and deliver response in blazing speed.

Why Apple

Up until this point there is nothing unique about this approach. Google, Facebook, Amazon and others are doing the same thing. What differentiates Apple from the others? In short, Apple’s good working relationship with 3rd party developers and companies.

Yes, Google has a lot of data, but the majority of the data is of low quality. For each category, there is almost always some company that owns datasets with higher granularity that are clearer and in some cases much larger. And there is no way that those companies are dumb enough to willingly partner with Google, after Google showcasing how it treats its partners. Facebook is in similar position. Instead of partnering with other companies, they rather acquire them and bully the rest of the market into playing under their own rules. The biggest issue though is that they don’t own the distribution platform which is in hand of Apple and Google.

Amazon is famous for the hostility towards their partners. And they also do not own the distribution platform (even they are making some remarkable progress with Echo).The rest of the market is missing more than just one piece of the puzzle to even compete.

This gives Apple an unprecedented opportunity to win the mobile search.

So how would the service work?

User talks to Siri. Siri translates the voice to a query and sends it to their cloud service. It then sends the query to the 3rd party service, waits for the answer and sends it back to Siri which will presents it back to the user.

Because Apple knows what apps each user installed on their phone, they could use this as a routing signal (to whom to send the query). But, it could also create a marketplace where others could bid to answer those questions. Very similar to how a stock markets operate today.

Imagine a user who wants to find the best pizzeria in her area. She has a Yelp installed on her phone. She asks Siri, which sends the query to the cloud where it is registered. Now other companies have 200–300ms to decide if they are interested to bid on this particular question. Lets assume Foursquare is interested. They would post a bid and if they win, they would be answering the question. Siri would now present the data back and mention, that it came from a different source.

It would encourage the user to download the app if she finds the response helpful. It would also allow Foursquare and others to build a business around it. They could make a deal with pizzerias in Manhattan to drive more traffic to them. Or if she is interested in a concert, companies could bid to provide the price for the ticket. Now each developer would get access to the whole user base instead of to the ones that actually installed their apps.

This kind of service would not be a small feat. Apple would have to fiercely focus on the speed (no user wants to wait 5–10s for an answer), figure out when the question is personal and requires some history of data (for instance if a user asks something like “What was the sushi restaurant I liked last time I was in Seattle?”) so it can route it to the right party, and more.


By talking to Siri, users would not feel like they’re giving away any personal information involuntarily. If you watched Google’s presentation of allo and duo then you probably know what I mean. Who really wants to send every single piece of text they write to Google’s servers? Even if you don’t care at all, do you really want a bot to post (or even suggest) whatever it feels like during your conversation?

To make my point more clear, by talking to Siri a user makes a conscious decision to share her data with Apple. This is a huge deal for Apple that in recent years focused more on users’ privacy. It’s also a big selling point for a growing portion of the concerned population.

Once Siri was good enough, Apple could venture it out to other devices, from cars to toasters. They do not need to keep it within the iOS boundaries and could reach this way even more people that are not their direct customers.