Enterprise chatbots: The helpful coworker your employees need

 In Solutions

It’s not hard to find negative stories about chatbots. You hear about them getting stumped or easily confused or harassed by users. You may even have your own story of a chatbot customer service interaction gone wrong. And granted, for consumer interactions, chatbots still need to make some progress.

But while it’s not yet fully baked for consumers, the technology is already finding applications in the enterprise, where chatbots are taking on the repetitive, mundane tasks that dog everyone from finance to HR to IT.

If you haven’t started exploring how chatbots can save you time, improve efficiency, and free up employees to handle more strategic projects, let this be your introduction to what’s possible.

The two types of chatbots that can help

Chatbots are primarily either informational or transactional.

As their name implies, informational chatbots provide answers to questions, dishing out information on anything from product details to analytics. Especially when there are frequently asked questions within your organization, chatbots can deliver an answer a lot faster than an email to a colleague, saving both employees time by drawing from questions that have been answered before.

Transactional chatbots complete tasks like booking hotel rooms, paying bills, or resetting passwords. These types of support functions can save your organization time, especially for repetitive tasks, like onboarding new employees. From filling out paperwork to sending handbooks or other onboarding materials, chatbots can lend a helping hand that takes some of the administrative burden off HR.

Chatbots in the enterprise aren’t entirely new. Intel has used an HR bot named Ivy since 2013 to answer payroll and benefits questions. Overstock.com has a chatbot named Mila that reports to managers and adjusts schedules when employees call in sick. Lloyds Bank uses a chatbot to help employees gather internal information from the organization’s entire knowledge base.

The use cases are as varied as the organizations putting bots to work.

How chatbots save time and effort

Natural language processing has come a long way. While bots could save time in the past by connecting employees and sharing information, natural language processing allows bots to better understand the intent of a question or request, adding even more efficiency.

For example, when an employee sent a help desk request to IT, the botnets of the past would only provide light automation. They might create a ticket or send the issue to a certain email address, but they couldn’t determine what type of request it was or do much more than copy and paste information.

Now, with natural language processing, chatbots can determine that classification, understanding the type and severity of the issue and creating a ticket along the way.

The same holds true for documents. If someone requests information, rather than sending the whole document, chatbots can supply the excerpt that’s relevant to the question. For doctors that can sometimes struggle to find notes in electronic medical records, a chatbot can deliver what they’re looking for so they can spend more time interacting with their patients.

The 3 most common challenges with chatbots

While bots offer potential time savings, efficiency, and automation, implementations don’t always go smoothly. This is why you’ve probably seen negative stories about chatbots out of control or just plain ineffective.

But most of these issues arise from a few common challenges that chatbots pose. Organizations that have had trouble with chatbots typically report inaccuracies, lack of flexibility, and their inability to interpret certain aspects of speech.

There are a few reasons why this might happen:

  1. Limited vocabulary: The chatbot doesn’t have an effective dictionary to reference, and gets confused or stumped about what a user is asking. Teach it the specialized terms, acronyms, and other internal lingo and vocabulary that’s second nature to employees.
  2. Limited visibility: If the developers haven’t clearly defined an organization’s major questions, a chatbot will have trouble fielding those questions. Take the time to sift through the many requests a department gets and you’ll find that there’s a core group of questions that get asked again and again in different forms.
  3. Limited knowledge: The bot doesn’t have enough information to fulfill a user request. Involve the correct knowledge workers early on to field questions and establish a strong base that the chatbot can pull from in the future.

Each of these challenges is easily corrected or, better yet, stemmed from the start. While you might hear negative stories about chatbots, most of them begin, at least in part, with these three challenges.

And while the steps above can help limit or eliminate issues in general, nothing beats employee feedback on their experience with the chatbot. You can take note of common pain points specific to your organization or bot, and ultimately improve the experience for everyone.

Bring on the bots

The technology behind bots will continue to improve, but by getting started now, you can begin to build out the knowledge base that chatbots need to effectively answer questions or perform tasks.

Every organization will have different needs. If you want to discuss your specific requests or explore how chatbots might help your organization, get in touch and we can continue the conversation.

Related Posts: You may also be interested in...

Leave a Comment

19 − 1 =

Pin It on Pinterest