The AI assistant space is getting more and more crowded very quickly. Whether it is Alexa, Otto or Amy, new names are being announced from both established and new tech players trying to define the post-mobile world. Impressive funding sums back this sector and there is a lot to gain back from technologies that automate everyday processes. With its not very applauded customer experience, there are a few lessons for the insurance industry to learn and innovations to keep an eye out for. Here is a review of the most active companies in the virtual personal assistants realm.
The original AI assistant, Siri responds to voice commands by talking to you. It has been embedded into iPhone, iPad, Apple TV, Watch, Apple CarPlay. Getting 2bn requests weekly, Siri is knowledgeable about a variety of topics from news and weather to directions and local businesses. It can send messages, place calls and make dinner reservations through OpenTable. Despite the hype around Siri, since its introduction in 2011, it has largely remained the same. Little improvements have been made, one of the biggest pitfalls being the fact that it cannot communicate with other apps and services such as Spotify and Slack. Apple, what are you waiting for?
Available on Amazon audio tools such as Amazon Echo and Fire TV boxes, Alexa is said to be making its way to other connected devices such as alarm clocks and pet feeders. This assistant uses an open API, so it can connect to other smart home gadgets and even place orders on Amazon by voice commands. As a disadvantage in comparison with Apple and Google virtual assistants, it is limited to household appliances since it can’t be integrated with smartphones.
For all of you out there who can’t let the “art” of texting go, Cortana responds to both voice- and text-based commands. Another plus is that it is supported by Windows, iOS and Android devices even though its performance on the latter two is not as good as on the former one. The list doesn’t stop there: it can track packages, hook into Windows apps, talk to Skype bots and even quote Shakespeare. Microsoft is looking to integrate it in the Office package. It will be interesting to see how that will turn out. Oh, and it also tells jokes.
Otto & Viv
Samsung’s prototype virtual assistant Otto launched in April 2016 resembles Amazon Echo, but unlike the retailer’s device it has got an HD camera that can stream video live to your phone or computer. Voice activated, it connects to the Internet and other smart appliances such as lights. Otto can answer questions, order products and play music on command. It can recognise faces and it even has a face of its own making it look like a little robot. However, just when we were getting excited about Otto, Samsung announced that its new Galaxy 8 smartphone will come with integrated AI assistant Viv after buying Viv Labs Inc in October 2016 (operated by Siri co-creator Dag Kittlaus). This is an undeniable move into the IoT realm because Viv is an open platform and supports third party services. This will not only allow Samsung to integrate Viv into its future lines of home and business appliances, but also allow the devices to communicate with one another. It is claimed to be capable of answering complex questions such as “Will it rain near the Big Ben after 7pm in three days?” Its natural language observed in demos has had the techies on the edge of their seats, however it remains to be seen if it delivers on expectations.
Google’s own version of a virtual PA is voice and text activated (depending on the device) and does what it is expected to do – remind you of important events, make reservations, play music and translate. It is embedded in Google’s latest gadget reminiscing Amazon Echo – Google Home. This Android assistant uses neural training and deep learning to interpret nuances in the human voice and identify human behaviour based on its variations. Maybe the most advanced and sophisticated one yet due to the fact that it understands everyday language, makes detailed recommendations and carries out important tasks such as booking a flight. Google’s objective with its assistant is to replace search and mobile apps as we know them.
No, not M from the James Bond series, but Facebook’s concierge service. Since only a small number of people have access to this on-going project, little is known about this PA. What differentiates Facebook’s assistant from others is the fact that people will collaborate with the machines to answer users’ questions. The company hopes that the AI-human interaction granted by employing thousands of human helpers will aid the digital assistant learn from real people and become more human in the end when it will work on its own. Facebook has made a big commitment to investing in chatbots earlier in 2016 and is now offering software developers tools and support in the hope of improving its chat assistant.
You might not have heard of this one, but it might come in handy when your search is very specific in nature such as supermarkets in a 3-mile radius that are not Tesco. Available on iOS and Android, it connects to third-party services such as Yelp and Uber. On the downside, it can’t hold long conversations and the app seems to be used only as a selling point for the company’s Houndify service which allows third parties to add voice to their own devices.
While not as functional as previous ones, Ozlo will make the foodie in you jump with joy. Or the busy professional who is always on the move. Why? Because it finds you places to eat and drink by internalising data from other services such as Yelp and Foursquare. It has a limited chat capacity, but it can answer follow-up questions about menus or opening times. It seems to be more of a chatbot, but its creators promised it would become bigger based on machine learning. Following a $14 million round of funding, Ozlo is now available on iOS and the web.
You’ve probably have already interacted with this one to schedule meetings. Created by a New York start-up called x.ai, it is a single-purpose assistant whose objective is to book meetings. You just copy in Amy’s email address (firstname.lastname@example.org) when sending out an email and the assistant takes over. It knows your diary inside out and manages communications with other parties on your behalf. It recognises everyday language and learns your preferences. The most obvious pitfall is that humans still need to verify the data.
Developed by Personal Health 360, Shae is a virtual health assistant. It collects data from more than 10,000 data points (initial questionnaire, family history, body type and even pollen count) to offer users highly personalised advice on living a healthy life. It connects to wearable devices to find out your mood and asks you questions if something is not in the normal parameters. For example, if your heart rates go up as an indicator of stress, Shae will tell you what to do to fix that. Shae is in private beta at the moment, with an estimated price for an annual subscription of $197. I think I see some applications in the health insurance sector for this one.
China’s leading search engine, Baidu, has entered the race for AI assistants by introducing Duer. The virtual helper completes tasks on your behalf, it doesn’t just give you instructions. For example it can order a meal from your favourite restaurant based on previous orders. It can book you places at your local cinema, favourite spa, gym or pet groomers. Baidu’s business model with Duer is different because it takes a cut of anything booked through the assistant.
Since they’re supposed to make our daily lives easier, virtual assistants could be the ones to get our insurance policies and claims sorted out. First of all, a virtual PA will improve user experience because customers will be talking to a familiar voice and will be able to do everything from the comfort of their couches. If integrated within an IoT ecosystem and tied-in with other smart devices, the assistant could evaluate the gadgets’ condition and alert the insurer of choice about a broken or stolen device. On the insurer side, access to customer data will enhance the accuracy of underwriting and risk assessment. Going even further, vocal interaction could offer deeper insights into human behaviour that companies have never had access to before, at least not on a large scale. Intonation can give away preferences and emotions.
Are you using any of the above AI assistants? Share your experience in the comment section below.
Marketing Assistant @Ingenin