AI supported virtual assistants will change the way we produce and consume news – and almost everything else we do in life

Brushing your teeth in the morning is about to become the most important moment of the professional day: your virtual assistant will brief you on news and events around town or the globe; it will highlight industry, market or relevant news it feels you need to know. Forget key word setting, or selecting news channels, it will find what you need to know culling thousands of sources; it will not only go over your agenda and manage messaging, it will prioritize according to what it has learned about you; in 15 minutes you will get done what even with today’s automation takes 10 times longer. The birthday flowers for your wife are arriving at work at precisely 1:30 PM in the afternoon when she returns from lunch with her friends and dinner reservations at your favorite restaurant have been confirmed for days!

So far it sounds very much like something that is nice to have and maybe even a cute novelty, so lets get busy and dig deeper. Your virtual assistant highlighted the new rates from the Federal Reserve and projected inflation for the year just in time for you to ask it to re-calculate all the numbers for your 8:30 AM management call and email them to all division managers. You will still be in the bathroom by the time your managers are receiving the new updated presentations.

The weather part of the morning briefing included the above normal winter storm currently hitting New York causing 2 hour delays in mass transit – it has already texted key people for your 8:30 AM meeting to plan ahead and confirming meeting. It is waiting for further updates on traffic to manage arrival at office and monitoring the big lunch meeting as well.

Still not impressed: the lab results from your latest visit to the doctor are in and your assistant will walk you through them step by step providing you with answers to any questions you may have based on the most up to date healthcare libraries and millions of patient instances. It will recommend steps you should take based on the evaluation and confirm these with your physician. Cholesterol high? Ask your assistant for a 5 year average and run against your demographics.

As far as your financial health is concerned, your virtual assistant will have more resources at its disposal than any broker can even hope to access. And in fact these will not be traditional static resources but rather dynamic-real time resources based on millions individuals investments, transactions and to the moment performance. And this is just the beginning on the  financial side, how about checking your credit card charges and producing your corporate expense reports – done!

Retail will never be the same. Take groceries for example, your assistant will know what you need, what you like and order it for you. It will pay and arrange for pickup or delivery when it knows you will be home. It will suggest new items and manage the whole process based on your lifestyle – having a dinner party and running low on wine? Not any more. Schedule the dinner and tell your assistant who is coming, it will have your Dad’s favorite drink on hand.

And maybe the biggest overall impact could come from your virtual assistant at work. Lets say you are a journalist and want to write an article, do you want to know the highest trending subjects? Just ask. Chose your subject and begin to ask questions using natural language, your virtual assistant uses NLP (Natural Language Processing) which means it works based on natural language queries, searches very large databases and answers in plain language. Your virtual assistant will use  archives going back decades for tens of thousands of sources instantly.

Work internationally and in other languages? Google has re-engineered Google Translate using artificial intelligence and now provides API integration for industrial purposes (large amounts of translations).  Ease of use is improved tremendously but more important is quality – by adding AI integration Google Translate has improved more than any machine translations ever had and now improves with every use. Excel, the financial spreadsheet we all know which up until now most of us use 2% of capabilities is now natural language  processing driven. Want to see cost of good sold adjusted for different estimated growth and adjusted for inflation across 1,500 products and 10 divisions? Just ask.


“OK Google, what planet am I on?” The Five Virtual Assistants, Compared

Siri, Cortana, Alexa, Google Assistant and Bixby…oh what fun it is to keep up with the ever-growing list of virtual assistants who claim they can improve your life, magically give you more free time to do things you love, and lower your cholesterol.

Or not.

We recently concluded the end of “Tech Announcement Season” for the first half of 2017, with Apple’s Worldwide Developer Conference grabbing the usual, final spot. And over these past few months, we’ve certainly heard many announcements about the assistant platforms from Apple, Google, Amazon, Microsoft and Samsung…and why consumers should choose their virtual assistants over others.

But what is it about these virtual assistants, really?

For Apple, it’s all about user privacy and trust. Siri’s smarts stay on your devices and your data stays within the Apple ecosystem. And with iOS 11, she will get smarter, and be able to surface more relevant information related to your day (and life) at the appropriate times. How that exactly works, Apple didn’t really elaborate on, but they encourage you to just trust them to make it work perfectly.

For Google, it’s largely the opposite. The search giant would love to connect every piece of your life into the Googleverse, and let the Google Assistant magically help you with as much as possible. And instead of the Assistant’s smarts staying on your device, they stay mostly in the Cloud. Your data is anonymized and whisked in with other data and requests from other users to create a better experience for everyone. And, you can be easily marketed to, wherever you might be in the Google ecosystem or on the web.

“The virtual assistant that is accessible to the most amount of people, can help you with the things that actually matter in your own life, can understand you most of the time, and that you inherently trust…will be the one that ultimately succeeds.”

For Amazon, well, ideally they’d love to sell you a few (more) things. Alexa would be more than happy to re-order some paper towels for you, or perhaps some jasmine rice for that Middle Eastern feast you’re making tonight. User privacy and trust is, well, somewhat assured. The assumption is that your data stays within Amazon, as they essentially want to market products to you, and they’re generally not sharing it with third parties. When Amazon launched Alexa with the Echo in late 2014, they took everyone by surprise. It made sense eventually, and they have used their own cloud infrastructure to make it possible. But building a smart assistant from scratch is no easy task.

For Microsoft, it’s about basic assistance with simple questions from users of the ever-ubiqutous (but rarely headline-making these days) Windows. While Cortana has the usual list of assistant capabilities like setting reminders, getting traffic information, finding sports scores, looking for local concerts you might like and similar tasks…she has never really made headlines with a standout feature. And only at this year’s Microsoft Build conference did the Redmond software company announce that Cortana would gain 3rd-party “skill capabilities”, similar to Alexa and Google Assistant. Microsoft is clearly playing catch-up here and working to stay somewhat relevant in the space.

For Samsung, it seems to be a “me too” strategy. (Though they certainly wouldn’t say that.) Bixby was rumored for months, and had been thought to potentially give Alexa and Google Assistant some fierce competition. But instead of integrating the excellent Viv context-aware assistant technology that Samsung bought in 2016, Bixby turned out to be a re-launch of its own S Voice service. There isn’t that much that’s known about Samsung’s data protection, cloud systems or truly how Bixby works. The South Korean conglomerate certainly doesn’t have the data-rich resources of Google, but Bixby is technically trying to compete with the Google Assistant. (On Android phones no less!) Bixby isn’t fully operational yet, but Samsung promises it will be soon.

Alright, so now you have some background. But before I go into some real comparisons, let me sum up this entire race, or game, or battle for you:

The virtual (or is it digital?) assistant that is accessible to the most amount of people, can help you with the things that actually matter in your own life, can understand you most of the time, and that you inherently trust…will be the one that ultimately succeeds.

The mission of all of the digital assistants is to provide us with relevant, useful information when we make an inquiry. Of course, some of them are better at this than others. But I thought that a very basic question would certainly be answered correctly by all of them, right?

“Which planet am I on?”

Sigh. I thought to myself, “aren’t these things actually supposed to be smart?

Both Siri and Cortana gave me a web search result for my question. (Not much of a surprise)

Google Home said “Sorry, I don’t know how to help with that.” (Really? GOOGLE?!)

Alexa initially said “Hmm, I’m not sure” for a few days. But eventually started responding “We are on planet Earth.”

(I couldn’t test with Bixby, because well, nevermind.)

I was completely shocked that the Google Assistant couldn’t figure it out, but pleasantly surprised that eventually Alexa figured it out. (Maybe there really are people analyzing these queries in real-time up in Seattle!)

Perhaps it wasn’t a fair test. Is that question really something that matters in my own life? Probably not, but what about an extraterrestrial who comes to visit 150 years from now? Anyways, my point is that even the seemingly-simple question that has an obvious answer demonstrates that these “assistants” really are not necessarily “intelligent” yet. They just know how to help with certain categories of queries, assuming they have the relevant and useful data somewhere that can be retrieved and relayed back to the person requesting it.

Let’s move on to the details of each platform, ranked from least likely to most likely to succeed:

Bixby (Samsung)

Launch date: April 2017
Platforms: Samsung Galaxy S8 (Android)
Has a smart speaker? Not yet
Knows which planet we live on? Probably not
Pros: New unique concept with contextual assistance with the backing of a large company
Cons: Bixby Voice is not yet operational, other services are currently very limited, no 3rd-party integrations yet, future is unknown

The newest assistant, and currently in last place, is poor little Bixby. Only Bixby Home and Bixby Vision work right now, essentially a set of Google Now-esque “cards” and some basic image recognition / shopping capabilities for the built-in camera and photos app. Bixby Voice is supposed to provide the general voice assistant capabilities like weather information, general knowledge and some local app control but that’s all right now. The U.S. launch of Bixby Voice is scheduled for late June, and it’s unknown what it will truly be capable of. Bixby is designed to also work within apps, to give a user context-sensitive help where they happened to be. For example “Hey Bixby, screenshot this and email it to Bob in Accounting…” The word “this” would refer to what’s currently on the display. However, only Samsung’s own apps will be supported when Bixby Voice goes live. Third-party app support will come “later”. Context in the digital assistant world is king, but if your assistant doesn’t do all that much to start with, could it ever wear the crown? And right now, Bixby only works on Samsung Galaxy S8 devices.

Samsung recently announced that Bixby would be coming to its Family Hub 2.0 refrigerators, with an update for those already shipped. Its plan is to eventually integrate Bixby into most all of its products, which will probably take longer than they anticipate. There’s been no word on whether Bixby will come to other platforms like iOS or Android. And, while there was definitely some talk and demos of how Bixby could interact with SmartThings hubs and devices in the smart home, there has been zero mention of a smart speaker. Of course, since Bixby Voice hasn’t launched yet, why launch a smart speaker? I really feel that Samsung should have teased Bixby with the Galaxy S8 launch and “turned it on” once it was actually “ready for action” months later, as in the holiday shopping season. Pair that official launch with a smart speaker that can easily tie into SmartThings and you’d have a good shot at competing with everyone else. But at this point, Samsung has a a long road ahead to implement Bixby across its product lines and expect to catch up to the rest of the players.

Cortana (Microsoft)

Launch date: April 2014
Platforms: Windows, Xbox, iOS, Android
Has a smart speaker? Coming soon (Invoke, HP smart speaker)
Knows which planet we live on? Nope
Pros: Integration with Microsoft services and hardware platforms, future Microsoft Home Hub with Cortana could be very powerful
Cons: Limited intelligence means minimal information available, few capabilities, no smart home integration yet

Cortana on Windows 10 desktop

Cortana on Windows 10 desktop

Ah, Cortana. There have been some news blips here and there about the Microsoft digital assistant, but it just hasn’t made the waves and major headlines that the others have. The challenge has been the platform: Windows. Originally Cortana launched on Windows 8.1, and of course, came with Windows 10. As Microsoft’s mobile strategy has been, well, challenging, it’s been very sparsely represented on Windows phones. So largely Cortana has been available on relatively new desktop, notebook and tablet devices from Microsoft, and the Xbox One. But here’s the other challenge, people tend to use their desktops and notebooks differently from phones and tablets. People tend to be more accustomed to using some form of voice control on their smartphone and/or tablet, but not necessarily on a full-sized computer. Microsoft’s Surface line blurs the line between notebook and tablet, but at least they have mics on-board vs traditional desktops. And, Windows just doesn’t have the “cool” factor as compared with Android and iOS.

Harman Kardon's "Invoke" Cortana speakers

Harman Kardon’s “Invoke” Cortana speakers

The reality is that not much development has been going on with Cortana that has truly notable. Although that could soon change. Cortana is will soon start having third-party “skills” made for it by developers, similar to Alexa’s set of skills that helps her do more things. And there have been two announcements about “smart speakers” that will work with Cortana recently, by Harman Kardon and HP. There’s no pricing yet for either of these speakers, but hopefully they’ll be priced less than the insanely-expensive HomePod that Apple just recently announced. While Microsoft says that over 400 million devices now run Windows 10, a few quick scans of discussion groups on Cortana say that not that many people are using it. Could that quickly change with a smart speaker launch? While it is indeed really natural to “talk to the room” vs a smartphone, tablet or PC, if relatively few people use Cortana now, what would compel them to buy a speaker that worked with Cortana? My guess is that it would have to be an impressive experience, at the right price point. Perhaps the HP speaker will less expensive than Harman Kardon’s product. We’ll know more later this year once the speakers launch.

Siri (Apple)

Launch date: October 2011
Platforms: iOS, macOS, tvOS, watchOS
Has a smart speaker? Coming soon (HomePod)
Knows which planet we live on? Nope
Pros: Seamless integration into all Apple platforms, learns about you without sacrificing privacy, HomeKit device control is easy.
Cons: Few third-party integrations (all controlled by Apple), no “Siri app store”, limited AI due to privacy protection, limited knowledge about the real world.

Siri became the darling of the tech industry in late 2011 with the launch of the iPhone 4s. While Siri could execute some basic tasks and even had context-sensitive replies, it became better-known for its funny and sarcastic responses. And, while Apple has improved Siri over the years with better speech recognition, more languages, wider functionality and more natural-sounding voices, she still feels limited. The challenge with Siri lies both with Apple’s “go it alone” strategy, and stance on user privacy and security. Overall, privacy and security are good things, but they create a problem for Siri to know enough about the world (in addition to knowing things like your schedule) to be truly helpful. Many users tend to complain that when they ask Siri a question that seems relatively basic, that she just retrieves a web search instead of providing the actual answer.

The App Store and Apple’s renowned devices are perfect for a closed ecosystem, and keeping things “all in the family”, but getting data about the rest of the world and making an intelligent assistant useful appear to still be quite challenging for Apple. As I’m writing this piece, Siri still can’t directly answer me for a simple question like, “What time does Chipotle close?” She just provides web search results. (Cortana does too, while Alexa and Google Assistant answer with exact hours and related information.)

At WWDC 2017 earlier this month, Apple announced some improvements to Siri, mainly around new ways that you could interact with app categories such as banking and task lists. And while Siri will also be improved to “surface” information when it thinks that you may need it, it is still all based on information that you give it and how it interfaces with your apps. Siri isn’t “out in the world” gathering general knowledge and trying to assist you based on how that knowledge intersects with your own personal life, it still remains a bit more reactive than proactive. Apple is working to make Siri more powerful and have its intelligence reside nearly entirely on your own personal Apple device(s) for privacy reasons, and this is an incredibly tall order. I personally think that they will ultimately run into a wall where there will be diminishing returns on improving Siri’s capabilities.

Apple HomePod in white

Apple HomePod in white

Also at this year’s WWDC, Apple teased out the HomePod, a pretty expensive “smart speaker” that is marketed as a high-end audio device first, and a “Siri Speaker” second. While some of Apple’s faithful will certainly buy it, it could be a tough sell without a less-expensive model. It’s compatible only with Apple’s AirPlay 2 technology, and Siri on the HomePod might only be able to play tracks from Apple Music or your own music library. So if you want to play music from other services like Spotify, you may have to AirPlay it from another Apple device…not a great experience. Apple announced some new music abilities for Siri (specifically for the HomePod) but nothing else. Sure, the basics of Siri (weather, news, messaging, podcasts) work on the HomePod, and HomeKit (of course), but no other upgrades to compete with Alexa and Google Assistant. By the way, it’s priced at $349 and ships in December. That’s the same price as the original Apple Watch Sport when it launched in 2015. I and others expect Apple to tell us quite a bit more about HomePod in the fall. If they don’t, then it’s going to just be a really great speaker, with a side of Siri.

Alexa (Amazon)

Launch date: November 2014
Platforms: iOS, Android, Fire OS, most Amazon Devices
Has a smart speaker? Yes (Amazon Echo devices)
Knows which planet we live on? Yes
Pros: Affordable entry-level option, wide capabilities with 10,000+ Alexa Skills, very good voice recognition, easy shopping by voice
Cons: Amazon is still inventing their AI from scratch, Alexa doesn’t know everything about the world

Amazon Echo Dot during an "Alexa Call"

Amazon Echo Dot during an “Alexa Call”

A project from Amazon’s secretive Lab126, the Echo (and Alexa) came out of nowhere in late 2014. The Fire Phone had become a major flop when the Echo launched to an invite-only crowd of Amazon Prime members. Everyone thought Amazon was nuts for building their own digital assistant technology, an artificial intelligence platform and voice recognition, certainly not an easy task when essentially starting from scratch. But Amazon became the first company to successfully create a product for the masses that lets people just “talk to the room”, just like in futuristic sci-fi shows and movies.

And they had a hit.

While Siri had sarcasm, Alexa could get it done. And she gained “skills”, as Amazon calls them, built by 3rd-party developers. Today, there are more than 10,000 skills for Alexa. In early 2016, Amazon introduced a less-expensive version of the Echo: The Echo Dot. And then they followed it up with an even less-expensive version six months later, and priced it ridiculously low for Black Friday. Anyone with $40 could bring Alexa into their home. And she doesn’t come with a monthly cost, either. The only requirements to make her work are Wi-Fi and a free Amazon account.

Alexa is now available in the Amazon shopping app on iOS

Alexa is now available in the Amazon shopping app on iOS

Streaming music, weather information, your daily “flash briefing”, flipping a coin, sports scores, playing games, ordering dinner via delivery, getting a ride somewhere, and of course, commanding your home’s connected devices…Amazon invites you to “just ask”. The e-commerce giant has embedded Alexa into its tablets, shopping app and Fire TV products. More recently Alexa has started showing up in home automation gadgets—most notably the new Ecobee4 smart thermostat—which can include a speaker and microphone array so that “Lady A” can speak and understand your requests from anywhere in your home.

While some people may think that artificial intelligence is an odd thing for Amazon to pursue, it’s something they’ve already been working on for years. According to Quartz, Jeff Bezos explained in his annual shareholder letter that “though less visible, much of the impact of machine learning will be of this type – quietly but meaningfully improving core operations.” Amazon already uses AI in its e-commerce business, especially with figuring out how to get packages into customers hands faster and more efficiently.

After recently announcing the new Echo Look and Echo Show devices, it’s crystal clear that Alexa is not an experiment for Amazon. It wants Alexa in every pocket, in every home, and to be embedded in as many devices as possible. What’s the precise future of Echo devices, and their namesake friendly digital assistant? You’ll have to “just ask” Jeff Bezos.

Google Assistant (Google)

Launch date: September 2016
Platforms: iOS, Android
Has a smart speaker? Yes (Google Home)
Knows which planet we live on? Nope
Pros: Excellent voice recognition, fast and helpful replies, seamless integration into Google services
Cons: Not yet as mature as Alexa with device and service capabilities, data privacy concerns persist with living in the Googleverse

Google Assistant on iOS

Google Assistant on iOS

At Google I/O 2012, the search giant introduced Google Now, the predecessor to today’s Google Assistant. Google Now could produce card-based information, but it was not conversational like as Apple’s Siri. But even before 2012, Google was already gaining experience with voice recognition with Google Voice Search. And, of course, the information giant has been using machine learning for years to make its Google Photos and others apps more useful with facial and object recognition, text-based content analysis and recognizing patterns for advertising purposes. While Google only more recently started talking about its artificial intelligence efforts and how they fit into Google’s grand vision for organizing the world’s information, it’s clear that AI has been a big part of their strategy for a long time.

After teasing its new Google Assistant mid-last year, the successor to Google Now debuted in Google Allo (a new messaging service), its new Pixel flagship smartphones and landed in the humble abode with Google Home. The Google Assistant debuted with an easy-going personality, very fast response times and could generally give you the information you were looking for or help you with a task. And, the Assistant on Google Home enabled voice command of Google’s own Chromecast devices, basically turning it into an voice-activated home entertainment controller…something the Echo still can’t do. Google knew it had to catch up to the wide variety of “skills” that Alexa had going for her, and over the past year, it did just that. The Assistant has expanded its abilities not just in the smart home arena, but with new voice-shopping tools, food ordering and messaging, to name a few.

Google Home with gray fabric base

Google Home with gray fabric base

This year, Google is continuing full steam ahead with the Assistant, and more specifically, artificial intelligence. AI was discussed extensively throughout this year’s Google I/O keynote and conference. Google wants to be everything for you, from your email to your photos to your smart home devices. Just give everything in your digital life over to Google and they promise that your life will be better. In my experience thus far, it is. Google Assistant just gets it done.

But to beat Amazon in the race for the home, Google needs to launch a competitor to the Echo Dot. It’s obvious that a low-priced device will get more people to buy it and place it (and potentially several of them) in their homes. With built-in control of Chromecast devices, the current Google Home already has nailed the streaming music and video side of things, giving it parity with the Bluetooth streaming and audio-out capabilities of the Echo Dot.

While many, many people have already purchased a “voice-first” device, there are tens…no, hundreds of millions more people who have not yet done so. The recent Voice Report from VoiceLabs tells us that most people will ultimately buy one or more of these devices from only one brand, so Google still has the opportunity to “steal Amazon’s thunder” with a less-expensive “Echo Dot Killer”. This year’s holiday shopping season will most likely be very, very competitive in this space.

The Virtual Crystal Ball

So what’s my personal prediction for who will win this race? Well, as much as I love the privacy of Apple and Siri, and the wide skill set of Alexa—sorry Cortana and Bixby—Google seems to know what I want and tends to understand me the best (and the fastest.) I personally use Google Assistant far more than Alexa or Siri, and I get excellent, sensible information even without connecting all of my personal services and accounts to Google. Most impressive.

If Google introduces a less-expensive version of the Google Home later this year to compete directly with Amazon’s Echo Dot, it will be game on. Apple will be at the high-priced end of the market with the HomePod, and Microsoft will be in an unknown position in the market with Cortana’s new smart speakers. Samsung’s Bixby probably won’t have a smart speaker until 2018, but they could pull off a surprise and have some flashy SmartThings integrations this fall.

At the time that I write these words, my money’s on Google.

Artificial Intelligence News