The Rise Of The Bots: 11 Ways Your Business Can Prepare
Business is ripe for a bot explosion. The foundational technologies are available, industry behemoths are fanning the fire, users are demanding better experiences, and companies are looking for new ways to optimize their financial performance. Are you ready?
![](https://eu-images.contentstack.com/v3/assets/blt69509c9116440be8/blt4d2425abf1fac43a/64cb3ddf0c60c1fd1ad6a5ee/robot-1339192_1280.jpg?width=700&auto=webp&quality=80&disable=upscale)
Enterprises commonly use bots to reduce customer service costs and improve responsiveness to customer requests. Many of the bots in use lack any real intelligence, although the landscape is changing rapidly with the help of artificial intelligence (AI) and machine learning. Major industry players, including Amazon, Facebook, Google, and Microsoft, are all jumping on the bot bandwagon.
"You can now plug AI, machine learning, and technology resources into your bots without owning that part off the technology stack," said Felix Rios, technology and innovation manager of the market research unit at managed analytics solution provider Ugam Solutions, in an interview. "[Meanwhile], as a society, we are becoming more comfortable with the concept of talking to machines. We also increasingly favor chatting over calling someone. All in all, it seems like the perfect storm for [an] explosion of bots to happen."
So, what does the rise of the bots really mean for your business?
Increasingly, bots are being used to automate tasks, speed search, and carry on conversations with humans. Of course, we've seen some epic failures, such as Microsoft's Twitter Bot, and we're likely to see more high-profile failures along the way. Even so, modern chatbots are designed to use natural language processing (NLP) as an interface, so users can accomplish tasks in less time that previously required and enjoy a natural-seeming user experience.
"Up until a few years ago, the way we would accomplish a task was to point, click, and hunt through menus to figure out how to express our intent to the system," said Matt Buck, engineer and cofounder of voice interface development company Voxable, in an interview. "Sometimes the easiest way to tell a computer what you want it to do is to [use] natural language."
[What would you do if you had someone shadowing you all day? Read Adventures in Pair Programming.]
Most natural language implementations still require considerable improvement, which is where machine learning and developer smarts enter the picture. Since NLP isn't perfect yet, organizations need to understand the limitations of an implementation.
Avoid putting blind faith in a bot's ability to mimic human-to-human communication. After all, it's difficult for a bot to understand the nuances of language, the contexts in which language is being used, and a human's emotional response to a particular stimulus.
"We define AI in three stages," said Anand Rao, chief technologist at Pricewaterhouse Coopers (PwC), in an interview. "Assisted Intelligence, Augmented Intelligence, and Autonomous Intelligence. Most of what we are seeing in enterprise or consumer applications today are in the Assisted Intelligence category. They help automate routine tasks and have not fundamentally defined the way we do those tasks."
While there's still much to be learned, bots are evolving rapidly. Here are 11 factors to consider as you prepare your enterprise for the Age Of The Bot.
(Cover image: iLexx/iStockphoto)
If chatbots can't accurately detect human emotions, they can't communicate with humans on a peer level. Similarly, if a chatbot is unable to detect our emotions in the moment, it is unable to adapt to an individual's behavior accordingly. The first version of Siri was one example. A recent Facebook Messenger bot developed by Imperson to promote the home release of the movie Zootopia was another. During the scheduled week it was live, the bot instantly answered messages set to it via its Facebook page. Its response to this author's query was completely irrelevant.
Intelligence will distinguish brands, so beware of mindless bots.
Brand image always matters. However, when automating customer service-related tasks, some businesses overlook the potential negative fallout caused by ineffective chatbots that fall short of customer expectations.
"As new interfaces emerge, and marketplaces and intermediaries emerge, brands need to shift their model so [the brand image] can be surfaced appropriately within these interfaces," said Scott Susskind, partner and CTO of sci-fi prototyping company SciFutures, in an interview.
"Brands that normally have people that deal directly with the public [have to ask whether] there will be AI that is able to distill down their brand and communication styles."
Jennifer Griffin, VP of content integrity and insights at user-generated content marketing software provider BazaarVoice is skeptical about the ability of chatbots to communicate in an interpersonal way. When executed badly, they can completely erode the customer's experience, she said in an interview.
Most bots focus on specific tasks. In addition to executing a single function or a narrow set of functions well, the limited scope enables the associated AI to develop deep domain knowledge that also may be useful to other bots and an orchestration layer. Some futurists and technologists foresee a virtual personal assistant with greater capabilities than those available today.
"Initially, personal assistants will be functional [for activities such as] calendaring, traffic, and weather. But then, as the algorithms get better, and as we allow them to access our biofeedback and biometrics, it will be more about mood, emotions, and positive stimulation [such as] what we should read, what entertainment we should watch, and who we should connect with," said Ari Popper, futurist and CEO of SciFutures, in an interview.
Identifying prospective customers and qualifying leads involves manual work that could be done quickly and effectively using bots. For example, Voxable developed a conversational interface through which salespeople can access information about their targeted leads.
"We're replacing traditional research. The way it's done manually is by googling or using a CRM tool," said Matt Buck of Voxable. "The advantage of using the bot, instead of a new tool or [GUI], is you can ask it to do something."
Extending the functionality of an enterprise application usually takes considerable development effort. Enterprise software company Unit4 claimed it found a way to extend functionality using a natural language interface and bots to simplify common business tasks, such as time sheets, expense reports, and travel plans.
"ERP systems are monolithic. It's one big structure that takes a long time to extend. By adding a bot into the mix, we can actually start creating new feature functions with natural language capabilities that don't have to be built into the application itself. They can actually have their own lifecycle," said Claus Jepsen, chief architect, global head of technology and innovation labs at Unit4.
Bot platform provider Kore allows companies to integrate a chatbot into ERP, CRM, supply chain, and other enterprise applications. It also has a natural language interface that enables users to email or text the bot. The bot is accessible through Facebook Messenger, Slack, HipChat, and other messaging apps. There are also SDKs that allow web and mobile app developers to put bots inside their apps.
The velocity of business is causing enterprises to automate more tasks and processes, including some forms of decision-making. The decision-making capabilities are typically rules-based at this time. However, the basis will change as bots become more intelligent.
"A particularly powerful domain for bots will be in helping to automate and support decisions in enterprises. This is because businesses often have really advanced and well-defined workflows, as well as financial incentive to add automation support for better precision, safety, and speed," said Nathan Wilson, CTO and cofounder of AI platform provider Nara Logics.
Most people have a love-hate relationship with search. Sometimes it surfaces useful information, sometimes the results are completely irrelevant. Keywords and key phrases have their limitations, which is why search intent is now being inferred -- sometimes awkwardly -- using recommendation engines.
"We've really reached the limits of traditional 'index and search' approaches. They're great for returning long lists of search results, but very bad at assembling and consolidating information," said Dan Adamson, founder and CEO of risk-focused AI and big data company OutsideIQ. "[Our] bots have been trained to think and act like an investigator or researcher, identifying connections and risks that would be time-consuming and very costly to discover manually."
Stationfy also uses bots for search. Its natural language interface enables users to have one-on-one conversations about their favorite sports, live events, results, and breaking news. Instead of using search engines or hashtags to find videos of the best moments in last night's game, a Stationfy bot can quickly find those moments. Similarly, the bot used by content discovery platform Outbrain has a recommendation engine at its core that enables users to discover content based on their interests. The company is expanding its beta following its pilot launch with CNN.
Chatbots could serve as viable interviewing engines for qualitative research, because, if they're designed correctly, they could administer intelligent and contextually relevant survey questions based on such factors as a participant's smartphone use, location, and time of day, according to Ugam's Felix Rios.
"Traditional market research surveys can take a long time and often get abandoned. A bot, on the other hand, could potentially administer survey questions throughout the day, depending on where the participant is or even what [the user] is doing with [her] phone," said Rios. It could also send reminders to share feedback when she leaves a store or restaurant."
The behavioral data captured would be very valuable. Companies could use it to better understand the behavior patterns of individuals and to improve the services their customers consume, on an automatic, individualized basis.
Not all bots are customer-facing. Many of them operate behind the scenes to improve cyber-security. Others are used to streamline business operations. Integration platform-as-a-service (PaaS) Built.io built bots for its own use to manage the work environment in the office.
"We've managed to save tens of thousands of dollars per head on heating and cooling costs in the office. Bots, in part, [also] regulate access to our meeting spaces," said Kurt Collins, Built.io's director of technology evangelism and partnerships.
Software and app discovery portal Softonic built a bot with the goal of providing an alternative user experience.
"While the bots [we] are building are helpful for enabling games and app discovery, bots don't replace other approaches," said Softonic VP of product Gino Micacchi. "They're merely another way of interacting with users and helping them achieve a goal."
Software and app discovery portal Softonic built a bot with the goal of providing an alternative user experience.
"While the bots [we] are building are helpful for enabling games and app discovery, bots don't replace other approaches," said Softonic VP of product Gino Micacchi. "They're merely another way of interacting with users and helping them achieve a goal."
-
About the Author(s)
You May Also Like