In Part 1 of this series, I talked a bit about the history of Enterprise Search (i.e., an internal SharePoint- or other-product-based search engine that crawls your company's body of content and people profiles). I explained why Google search-result relevance is so superior to anything a single organization can provide its users, and how result tuning and other analysis work is so important to getting results users find satisfactory. Here in Part 2, I'm discussing an experience and results that could be possible with a new approach: making finding content, people, and even business processes easier for users by making them conversational…with your own company chatbot.
In the future (at least according to Star Trek), we'll all be having conversations with our computers to get anything done. With rudimentary speech-recognition software, we are already getting what we want from Alexa or Google using our voices—in a home/consumer scenario, anyway.
But what is the primary means of communication between people right now? If you go by volume of messages, the winner is not voice: really, who makes phone calls anymore? No, by far the leader is text/SMS or instant messaging—which means, essentially, that everyone is typing at each other.
Enter the chatbot. A chatbot makes interacting with a computer feel like you're chatting with a live person. I'm sure you've seen a consumer bot on a shopping website, one that intrusively pops up and asks if you need help, as if there were live people watching the site at all times, ready to offer a hand. An organizational bot can also help users find their way around a site (in your case, your intranet), but they can be tailored to do a lot more.
Built into Microsoft Teams is the rudimentary Power Virtual Agents, which allows organizations to build quick, no-code chatbots that can answer a list of questions. But beyond that, Slater Hill or another partner can use the Microsoft Bot Framework, among other tools, to build you a chatbot that is familiar with the content and even the tools in your organization's digital workspace.
More than just guiding users to a web page based on keywords the user enters, a bot can deliver documents or even people (in the form of contact cards) directly to the chat interface. It can also detect a user's intent to start a process, say, requesting time off, and ask a few questions to submit the request on the user's behalf. Now we're talking about a personal digital assistant for every employee in the organization.
A consumer bot on a public website analyzes the keywords in a user's query and does simple matching to locate one or more suggested links—basically a Search. But a tailored corporate bot has been programmed to recognize the names of customers, suppliers, cases/matters or processes (let's call those "entities") and continue to interact with the user to move the conversation along. Depending on the type of entity found, the bot can provide different sets of type-able options.
For example, if I type a question to the bot that includes the name "Tom Hill," the bot will recognize an employee and present a contact card, in the context of the chat, with all of Tom Hill's information. By clicking on, for example, "Start chat," or by typing the reply "chat" (i.e., keeping it conversational), I can begin a 1:1 call on the spot. Or I can type "files," and the chatbot will display a list of recent Office documents shared with me by Tom.
Now, employee lookup alone doesn't seem to add a huge amount of value (in terms of time savings) all on its own. But when I am able to find any person, or document, or SharePoint Site or anything in my Microsoft 365 tenant using the same chat interface, like a digital personal assistant, the savings start to add up.
Of course, if the question I ask the bot doesn’t have any recognized "entities," it can fall back on enterprise search and show me a standard set of search results.
This is where the value really starts to appear. In addition to entities, the bot can parse my typed query to look for available business processes or workflows I can kick off. For example, I can type "time off" and, like the Clippy of old (but without the cutesiness), the bot will process my intent to start the company Time Off workflow.
And rather than redirect me to the initiation form and let me figure it out on my own, the bot can simply ask me questions (start date, end date, reason for leave, etc.) and kick off the process for me. Now the bot is really acting like a personal concierge—especially useful for new employees.
Of course, my company's Time Off workflow has been written to send me updates through the bot to keep me posted on the progress of my submission. The same can be configured for other business processes as well, such as getting an expense item approved, kicking off a new contract process, creating a Microsoft Team or SharePoint Site, and more.
An organization chatbot can be launched as a version 1, with an initial set of responses and processes, and then trained over time to respond to many more. If your organization is not currently using Teams, Slater Hill can work with you to get it implemented securely to all your employees and partners. Once your people are accustomed to the experience of using the Teams client on desktop and mobile, our team of consultants will use the Microsoft Bot Framework and the Teams Developer Studio to build you a tailored, trained chatbot in three to six months, depending on the complexity and volume of potential responses.
Ready to start exploring what it will take to make your Digital Workspace accessible to users in a friendly, conversational way? Contact us with any questions you have.