Apple's artificially intelligent assistant Siri isn't exactly the sharpest knife in the drawer. But there is one specific piece of technological jujitsu Apple can perform that just might be able to fix that.

AI is hard.

It requires expertise, complex algorithms, and huge amounts of training data. With over $250 billion dollars of cash on hand, Apple can buy the first two.

The third piece, however, is especially hard for Apple.

The result is that Apple is barely clinging to third place in a recent " AI IQ" test, answering only 22 percent of 5,000 test questions of factual knowledge, compared to Google Assistant's 68 percent and Cortana's 57 percent. (Cortana is Microsoft's AI agent; Amazon's Alexa is in fourth, at 21 percent.) While in actual daily use, Siri answers questions and completes actions at a much higher rate, there's no doubt that thanks to decade-long investments in search -- and therefore answers -- Google and Microsoft are far ahead.

This is precisely the reason that when Apple unveiled HomePod, its smart speaker answer to Amazon Echo and Google Home, the company was forced to stress sound quality, not intelligence, as a key buyer benefit.

(As if Apple has ever wanted to enter the home audio market.)

The problem for Apple is that artificial intelligence is the foundation stone of competition between the modern titans of tech, including Google, Facebook, Amazon, IBM, and Microsoft. The titan that wins AI will be smarter, solve customers' problems faster, connect brands and consumers better, help people manage their lives easier.

And data is the new oil, not only of the new economy, but of the artificial intelligence race.

That's a problem for Apple, because one of its core selling points is customer privacy. While Google and Facebook vacuum data from our digital lives like the Saudis suck oil from desert sands and get smarter with every query, every like or status update, Apple promises to protect customer privacy and not use our data as fuel for competitive advantage.

There is one thing Apple could do to get smarter. And that one thing is central to how Apple became the company it is today.


Siri may not know if a friend of yours lives in Seattle, but Facebook does. Siri may not know what the top tech news of the day is, but Flipboard does. Siri may not know that your favorite afternoon drink is Iced Vanilla Latte, but the Starbucks app does.

You get the point: if Siri can deepen its integration into the very best of the 2.2 million apps currently available on the App Store, Siri might not have to be as smart as Cortana or Google Assistant.

Maybe, with a little help from friends, Siri could be even smarter than them.

And, Apple might be able to take this farther than any AI assistant currently can accomplish.

For example, Siri may not know which fridge will fit in your kitchen, but IKEA's coming augmented reality app might. And Siri might not know what vacation would be best for you, but checking the Calendar app and combining that with an email and Slack analysis might give Siri some clues about how busy you've been and if you need to crash on a beach or enjoy a more active holiday.

The opportunity is simple: use Siri to interconnect certainly key but maybe all apps, depend on partners to know things Apple doesn't, and return appropriate answers and action in response to Apple customers,

There are many details to work out in that kind of strategy, and technology to build. Apple would have to develop proper means of crediting app publishers and monetizing their contributions.

But company strategy needs to proceed from the strengths you have.

And this might just be Apple's best shot.

Published on: Jun 9, 2017