What are AI Assistants, and how can they help you be more productive?

Hey, Alexa, what's happening today?

If you reserve your first greeting of the day for your personal AI assistant- Alexa, Siri, or Google Assistant-you are in good company. In 2019, 45% of American households already owned a smart speaker or smart home device, and another 26 planned to purchase one soon. By 2020, there were 4.2 billion digital voice assistants in use worldwide. Voice recognition has become an expected and integral part of smartphones, computers, and even smart coffee makers. We ask our smart assistants to turn up the thermostat, turn on the lights, and tell us which way to drive when we're going somewhere unfamiliar. 

The speed at which AI assistants are becoming mainstream may be surprising, but there is an obvious reason why. They are good at what they do and help us multitask at a rate that previous generations could only dream of. This speech recognition technology fits seamlessly into our daily routines, making life that much easier. Personal artificial intelligence assistants can match the user's intent using voice commands, for a hands-free google search experience. They help you set timers, play music, and check the weather. Can't remember how much water you need for boiling a cup of rice? Ask Siri. Need a plumber that can deal with your clogged sink? Google Assistant has your back.

While personal AI assistants are the perfect answer-man for the hundreds of questions that come up during the day, they become less helpful as they get out of the house and clock in on the job. Siri is a bit out of her league when you're trying to analyze the information you've gleaned from half a dozen calls and interviews, and Google Assistant would be hopeless if you asked her to summarize that complex virtual meeting you just attended. Something else is needed here; an a.i. assistant with features specifically designed for your calls and meetings is required if you want it to join you in the workplace. 

In 2017 Amazon launched Alexa business, a newly expanded version of their well-loved personal AI virtual assistant meant to fill the gap at the office. Alexa Business is designed to be every employee's helpful aide, covering such tasks as booking conference rooms, scheduling meetings, and re-ordering supplies. Connections to a smart display gave the opportunity to visually confirm key decisions or review emails, and, if given appropriate commands, the smart assistant performed flawlessly. There might be room for improvement in flexibility and smoothness. With integrations from Salesforce, Concur, and Polycom, Alexa Business promised well, but it is still not yet mainstream four years later. 

Microsoft's Semantic Teams unveiled a vision for Cortana in which their digital assistant was the helpful, responsive, and capable subordinate everyone wished they had at the office. In their unveiling, Cortana reschedules meetings, books rooms, and sends off emails, all through a natural conversational back-and-forth. The voice technology is much improved, and the only thing that tells you there's no actual human on the other side of the line is the overly smooth efficiency. It's still just a vision, but once Cortana can do all that, there won't be any problem finding users.

Microsoft wasn't the first to envision a responsive, capable scheduling assistant for enterprise and business use. X.ai was a workplace tool designed to insert itself into a conversation and find meeting times that worked for both parties. Invisible software, it worked simply in the background-cc'd into conversations-and sent off meeting invites when appropriate. While X.ai had its fans, it was retired after an acquisition by the event management software Bizzabo. 

A new reiteration of the same idea is Evie, a scheduling coordinator put out by Siemens that, according to the reviews, does a great job scheduling around everyone's calendar. Scheduling genies are helpful, and it's nice to be able to have an invisible assistant there at your beck, ready to organize your workday. In general, though, we're seeing a trend where "scheduling assistants" are less AI-based and more of a generalized tool for suggesting that empty slot on your calendar.

After all, the best scheduling experience doesn't require machine learning when it comes down to it. Some basic functionality-- mapping an appointment request to an empty place on the time grid-- is all that is required, and the rest is an intuitive user interface. 

Tools like Calendly (which boasts 10,000,000 users) are great for general meeting booking, while solutions like Drift and HubSpot make booking sales meetings and software demos easy for websites. These tools take the work out of organizing schedules but don't lay claim to many special AI powers.

Other companies take a hybrid approach, offering services based on AI but keeping a human in the loop to help out at sticky points. This enables them to extend the capabilities of their systems far beyond the capabilities of their machine learning systems, and a (limited) human workforce can manage a large number of requests. In many ways, it's a win-win situation, both in terms of reach, cost, and capability.

If you've visited the website of your electricity service, internet provider, or even your local library, you're likely to have encountered another type of AI assistance that is almost omnipresent today: the chatbot. These are the chat windows that pop up in the bottom right corner of your screen when you're browsing a service-oriented website. These chatbots can automate support questions, book demos for your team, or pull up requested information from an internal database. Natural language understanding, trained on thousands of support requests, allows these customer engagement bots to parse misspellings, abbreviations, and the many creative variations of simple support requests. This natural language processing is crucial to creating a good customer experience, unlike the "select 1" of robocalls.

Nothing frustrates a customer more than to feel they are talking to a machine, so the developers of these bots work hard to incorporate natural language and mimic real-life chat support. A well-designed chatbot will leave you feeling almost as if you've been talking to a person. They also keep humans in the loop, giving customer support staff the ability to step in when the support bot can't answer a question. The better the AI technology behind the chatbot, the less often this happens--- but, even with a perfect system, there will always be that one person who doesn't hesitate to demand, "I want to speak with a human!"

AI schedulers are great; chatbots are helpful too. But what about an AI assistant that actually lessens your workload and simplifies the tasks you need to do on a regular basis? What about an AI assistant that would not only take record meetings, take notes, and send reminders but could actually analyze conversations, summarize key points and highlight actionable insights?

It's a promising idea- and, as it turns out, it's more than an idea. Smart meeting assistants have become a reality, and they're surprisingly easy to implement. Your calls and meetings can be automatically recorded, transcribed, and converted into a Wikipedia-style knowledge base that can be accessed by anyone you choose. Conversational analytics and semantic analysis allow the raw data to be converted into structured, easy to skim write-ups that summarize the meeting, and pull out action items. Refer back to key points in the conversation, explore topics discussed and decisions reached, and use a search function to locate key mentions over multiple meetings. 

The technology behind the automatic note taker is one step further than that of your run-of-the-mill virtual assistant. Not only does the AI note taking assistant use speech recognition to transcribe the meeting and natural language processing to make sense of the conversation, but it also goes on to analyze the data using deep level machine learning models that get at the heart of the discussion—and what you're likely to want to do with it next

What is the future of AI assistants, and what can we expect to see next? 

1. First, we'll start to see A.I. assistants get more competent and match our intent with higher precision as larger, more sophisticated machine learning models get incorporated into the assistant engine. 

2. Next, you'll have more stateful interactions with assistants. Instead of one-time interactions like "what's the weather," our assistants remember what happened yesterday or last week and provide more thoughtful recommendations to keep us on track. 

3. We'll also see more real-time interactions with an AI assistant on our meetings and calls. Hosting an AI meeting on Zoom where the AI note taker joins and records silently is becoming more common, soon we'll see note takers post real-time transcripts and reactions during the conversation. 

4. Combining stateful and real-time interactions open up a plethora of use cases and opportunities. This is especially true in the workplace, where employees are inundated with data, meetings, tasks, and follow-ups. This additional functionality may be the missing piece that will bridge the gap from the personal assistant experience in your home to the assistant for your workplace. While most people will want a clear separation from work and personal, similar to our personal and work email, you'll start to notice a convergence of functionality and user patterns.

We don't have to wait ten years, or the next wave of features to take advantage of all that artificial intelligence and notetaker software has to offer. AI assistants for your business have already joined in the action.  Providing summarized meetings, searchability, and collaboration across your hybrid workforce. If you haven't incorporated them into your daily routine at the office, you'll want to start now--- they can leave you more productive, better informed, and with more free time to work on what you care about.