I was perusing the web today and found an article on IT News entitled “How To Protect Chatbot Data And User Privacy” by Terena Bell. We are always focused on educating business owners and managers on securing their companies’ data and Equifax reminded us of the importance of security controls. Their massive data breach exposed 143 million Americans’ information and ultimately led to CEO Richard Smith’s resignation. You may have caught on from the title of this article that it is not about the security of these bots, but rather an overview of what the bots can offer and why you should seriously consider utilizing this emerging technology. Any new technology will have security requirements and precautions that will be necessary to ensure that your data is secured, so I would recommend Ms. Bell’s article if you would like to learn more about bot security. Before you start implementing bots in your company have a discussion with your IT provider about what you need to do to ensure that the information collected by these bots is kept safe and secured.
What are Bots
Bots are a form of software as a service (SAAS) that allow users to interact via a chat interface. Bots can be programmed to operate on a number of messaging applications including standard SMS, Slack, Facebook Messenger, and Telegram. Great bot design focuses on creating a user experience that mimics a conversation with a human.
Bots fall into two major categories:
- Those that are powered by rules
- Those that are powered by machine learning (ML) to achieve artificial intelligence (AI)
What is the difference?
Bots that are powered by rules are less robust than their AI counterparts. They operate according to a set script and will not understand if the user deviates from that script. That script is usually very narrow and so, for example, the bot may understand “Book me a flight to Paris tomorrow” but not, “What Paris flights are available tomorrow?”. Furthermore, no matter how many times you tell it what you mean, these bots will not “learn” a command or question that is not part of its originally programmed rules. I would argue that these are not the bots of the future, but they are a stepping stone in the right direction and an excellent solution for any business that wants to get a bot in the game without spending a ton money.
The bots of the future are able to “learn” from their interactions with users through ML. These bots take more time to create since the code behind them if far more complex than standard if-then statements. Just to clarify a common misconception: Artifical Intelligence (AI) is the concept that machines should be able to complete tasks in a “smart” way while Machine Learning (ML) is the current mechanism that programmers are utilizing to obtain AI. Forbes has a great article on the distinction between the two and a little history on ML and AI. Remember that the “smart” behavior of these machines is merely a self-correcting algorithm attempting to make its code as “least wrong” as possible.
If you have ever used Cortana, Google Home, or Siri then you have surely heard each AI’s respective version of “I don’t understand”. I think that everyone is in agreement that we are nowhere near a perfect AI. That being said, you can help these machines learn. If you’re looking for some helpful tips How-To-Geek has a great article on How to Train Siri, Cortana, and Google to Understand Your Voice Better. The article takes you through the steps in accessing the machines learning protocols. With Cortana, you can go into the settings and repeat a series of phrases to help her understand your voice commands. Google Home, on the other hand, is constantly sending feedback to the “mothership” and getting intelligence pushed to the device from this central repository.
The Future of Business Communications: Bots
It wasn’t so long ago that the best way to get information was to pick up the phone and dial 211. Auburn University’s James E. Foy Information Desk has been answering caller’s questions 24/7 since the 1950’s, and if you don’t believe me just call (334) 844-4244 to ask them if it’s true! Most businesses transitioned away from these manned call centers as the internet allowed companies to publish their information on a platform that could be easily accessed from anywhere. Email and online forms streamlined communications and facilitated the triage of customer needs. Simultaneously phone system call trees became easier to program and use so the demand for resource-intensive call centers has been slowly diminishing. Now we face the next great business communication evolution: Bots with AI.
Companies are using bots to assist their HR departments in answering questions that do not require organic intelligence or login, such as “How many vacation days do I have left?”. In the past, someone from HR would have to go searching in a spreadsheet or employee file for that information, but a bot like Talla can quickly query a database and have an answer for the employee in seconds of the question being asked.
Need to get feedback from employees but don’t have the time to put together a poll? Polly is the Slack poll and survey assistant that will create the poll, solicit and track responses, and then analyze the data for you.
Let’s say that you get an email from someone requesting a meeting, but you are swamped and can’t coordinate the time and date. Meet Amy: the AI scheduling bot that works with your calendar to schedule a meeting between you and your guest. She is pretty accessible, just cc: firstname.lastname@example.org when you need her help.
The external applications of Bots are practically endless. It all comes down to what you want to get out of a bot and how much you are willing to invest in developing the AI or rules that power it.
Bots can augment your existing sales and marketing staff. Imagine a single bot that is able to simultaneously help a hundred customers browse your inventory, answer questions, and make recommendations. Many of the biggest retailers have already developed bots to help customers find clothing and shoes that match their style and price range. On Facebook Messenger, ShoeDazzle personalizes the online shopping experience to give users the feel of a classic boutique and reaches out to users monthly to share new products and offers. The Kohl’s Bot can help users find retail locations and hunt down the best deals.
Reservations no longer need to be made through an antiquated online form when you can just text your bot the details of your trip. Marriott and Hyatt have both launch bots that will help you find and book hotel rooms, link to your reward accounts, and share articles and information with users. No doubt that this same idea will be (or has been) translated into all types of reservation systems including those used by restaurants and stylists.
There are bots that will help you shop for a car or find a home that meets your needs, diagnose car problems and help you fix them, and even hunt down a job that is perfect for you! Domino’s Pizza has a bot that will take your order and track its progress all through a series of written communications in a Q&A format.
What to Expect in the Future
Matt Schlicht, the founder of Chatbots Magazine and CEO of Octane AI, predicts that Bots Will Completely Kill Websites and Mobile Apps. In his article, he gives three very compelling reasons that bots will overtake what we consider to be the pillars of communication in the business world.
- Social messaging apps have become the #1 way that people are communicating, so businesses need a strategy to reach people on these platforms. Schlicht’s conclusion: Every business will have a bot.
- Bots are lean and fast – faster than websites and mobile apps. People like fast. Bots do not need to be downloaded, so they don’t take up valuable space on a user’s device.
- Bots are easy to use! They are designed to work with a user’s vocal or written commands. Since language this is the most natural way people are used to communicating, no button-based user interface can compete with a well-designed bot.
Mr. Schlicht may be right in his prediction. A static website will never be able to compete with a bot that learns and grows with the business as a result of customer interactions. The appeal of a mobile app goes away if an easier and faster solution is available without the hassle of downloading. Taking all of that into consideration, the bot revolution will not happen overnight. Even in a time of rapid technological advancement, the majority of businesses will hold on to their websites and mobile apps. Psychologists might suggest that the sunk-cost bias will prevent companies from ditching the domain while others would err on the side of caution in case this whole bot thing doesn’t pan out. Users are always slow to change because they like what they know. Many will forego bots in the same way that many consumers never left the flip phone behind. So, don’t go auctioning off your domain to pay for your bot, but keep in mind that business communication is going into an ever-more interactive space and hum-drum repositories of information will no longer be sufficient to maximize revenue or customer retention.