It’s a fact: Technology is always evolving, whether it’s the shift from desktop to mobile, the increasing use of APIs, or service-oriented architectures (SOAs) built entirely in the cloud. So, what’s one of the next big shifts?
Popping up more and more is the use of chatbots to streamline how we communicate with customers—a trend that’s directly correlated with the rise in messaging platforms, and an answer to an increase in consumer “app fatigue.” Bots have been around the web for years, doing things like automating tasks and reporting sensor data, but this is a new variety—one that’s riding the wave of messaging platforms like Facebook, WhatsApp, and Slack.
Bottom line: Businesses have to be where customers are, and that’s less inside mobile apps, and more so in messaging platforms. These savvy customers expect a seamless, real-time, customized way to get answers to their questions or find products they’re looking for, and bots that are integrated into messaging platforms provide an easy, cost-effective way to do this.
So what are bots, what’s the technology behind them, and is a bot right for your business? Let’s take a look at the evolving world of bots and see if you should consider developing one.
Bots are here—and here to stay
Bots are software and scripts that are designed to automate certain tasks, make searches faster and easier, and, to communicate with humans. This last part—talking to customers—is the job of “chatbots,” bots that are designed to carry out specific tasks then return the data, all while essentially “acting” like a human.
Bots all over the web perform myriad functions—from data scraping and web scraping, to scheduling meetings or acting like e-commerce concierges. Facebook has brought business bots into its Messenger platform; Slack has in-line bots for all kinds of tasks; a CNN bot will summarize stories for you; Operator will search products based on price, color, style, and more; and Twitter has enabled deep links that let companies redirect customers to Direct Messages for private conversations. Kik, a messenger service, is also open to bots.
Fighting app fatigue
With people downloading fewer apps, savvy companies are moving their marketing and services where their customers are via integrations with messaging platforms.
It may seem counterintuitive at first. Many companies spend a lot of time and money driving traffic to their apps, but with bots, companies can fairly easily leverage someone else’s app. Facebook Messenger has significantly more users than an airline’s mobile app or an e-commerce app, for instance. So, it makes sense for many companies to capitalize on a platform with a huge audience already. For an airline, their app may be used so infrequently that using Messenger as a way for customers to check into flights could be a winning strategy.
Around-the-clock customer service with bots
When it comes to customer service, having a chat bot (or combination of a bot and human customer service agent) handle messaging with customers is not only convenient and effective, it’s often much less expensive than a call center interaction. And bots are also very patient—they can wait if a customer leaves a conversation for an hour, a day, or a week.
Chat bots are automated programs that interact with customers over messaging platforms to answer questions, provide suggestions, and more. More and more companies are investing in bot strategies as a way to help them augment customer service, provide better service, and save money in the process. Even fast food chain Taco Bell created a Slack bot with artificial intelligence and natural language processing that can take orders and payments for food.
It’s not actually a bot strategy—it’s a customer engagement strategy
How can having a bot improve your customer’s experience? In an era when anyone can pick up a smartphone and search for what they need, how will a bot provide next-level service for users? By pairing the two best aspects of bots: engagement and automation.
First and foremost, a bot should be part of your overall customer engagement strategy—or maybe even just a part of a single marketing campaign to support that strategy, like a bot that lets customers find the right running shoes for them to coincide with a company’s launch of a new shoe design.
Facebook’s bots on its Messenger platform have been a groundbreaking change in customer service, e-commerce, and even how users search the web for information. But an important question to ask—as with the adoption of any new technology—is whether bots are really adding value or if they’re more of a deterrent. Do customers like the convenience of bots, or are we just adding more noise to an already highly connected world?
For many companies, bots are about connectivity. That means communicating with your customers where they’re already spending most of their time. Start by looking at app usage: What apps do your customers use the most? Your own apps, or apps like Facebook Messenger? The answer is often overwhelmingly the latter, so that’s where many companies are looking to place Messenger bots.
It’s also about response time. Customers already know how to get things with the click of a button and expect that immediacy without having to wait for a person on the other end of the line—whether they’re shopping, booking travel, or checking their bank accounts. Thanks to bots, customers get the answers they want quickly without having to wait for an email reply. And if they need to step away and come back later, the bot will be waiting.
What’s the technology behind bots?
Bots are automated software with a human-to-machine interface that responds to queries by pulling data from a database and returning it with a conversational tone.
Many mobile and web-based bots are a mix of:
- JSON data
- APIs and HTTPS connections to interface with the server
- The cloud, which is the storage component of your bot, the size of which will depend on things like how long you opt to keep messages
Bots that seek to be more “human” involve more complicated (and expensive) technology, like
- Artificial intelligence (AI), with three degrees of complexity:
- Assisted intelligence—automation with routine tasks
- Augmented intelligence
- Autonomous intelligence
- Sentiment analysis
- Natural language processing (NLP), which helps bots distinguish between literal meaning and slang meanings, and acts as the actual interface through which people interact with the tech
- Machine learning—which helps to improve bots’ performance along the way
Fortunately, the tech barrier to entry for businesses that want to try out a bot is not high. The technology that creates the foundation of bots is readily available and accessible for developers skilled in back-end development, and can be integrated without much change to your existing technology stack.
The best bots will be hyper-specific bots
How do you implement bots without them becoming annoying? Making conversation meaningful and not overly “chatty” will be key.
Chat bots present some obvious challenges and barriers to being a success with customers: They have to be conversational, which entails extensive programming, AI, machine learning, and access to lots of data. In this respect, the more simple and focused a bot is, the better. With fewer questions to answer or predetermined responses, the bot has fewer chances for error and narrowing the data you’ll need to provide the right answers.
Another way to sidestep complications with human-to-bot conversation is by building your bot with a “smart keyboard.” Instead of inputting a question to the bot, a menu-style keyboard can provide users with pre-selected choices of follow-up queries to send back to the bot. This works well if your bot has just a few options for how it can respond (about 5-6). In this way, bot interactions can be engineered to feel like a lot like web pages with custom menus, location data, forms, and more.
Considerations with bots
Having a man-machine interface for conversations does have its risks. Here are a few things to consider and look out for:
- Companies will naturally be hesitant to use bots because a bot will be a voice for the brand, and it can’t be monitored at all times.
- Users may get frustrated if a bot can’t answer their questions, gets confused too often, or is slow to respond.
- Without AI in place, automated responses could unintentionally offend someone, be robotic in response to a genuine issue, or unable to answer the question at all.
- If a bot doesn’t have the AI to respond well, it’s probably best to make the bot as simple as possible. It’s also important to keep the bot’s software updated, so rather than taking on a bot for the long haul, you could use it as a more temporary solution (or as a supporting aspect of a marketing campaign) to test the waters.
- When it doubt, build your bot so a human CS agent can jump in and take over if the bot is struggling.
- Unfortunately, Facebook Messenger doesn’t provide much insight into the performance of your bot, so it’ll be difficult to know how it’s doing.
Ready to bring a bot on board?
Although bots have been around for a long time, they’re getting smarter thanks to all the data we have and to advances in machine learning. Yet, the best bots will probably be the simple bots—for now.
Can a bot build on some existing functionality of your app, and make it easier for your customer to use? Or, will it streamline customer service by automating simple tasks? While there are many benefits to bots, not all bots are going to be worth the time and money. When considering options, the decision should be less about the bot and more about how you’re engaging your customers.