HR portal and call center waiting loops are over! With our new Questions and Answers Chatbot HR goes digital, moving away from aging FAQs and Wikis and call center experience to providing an interactive, easy to use knowledge base to your employees answering your people’s most frequent questions.
In our second post on the topic of chatbots let us look at another standard scenario - the FAQ bot.
The problem statement at hand is a straightforward one and has been solved through different means in the HR service delivery domain since the begining of time. With the many questions our employees are asking every day, do we have queries which repeat themselves and if yes would it make sense to collect those and present the answers to the customer in an attractive and informative manner so that we can focus our attention on other, potentially more strategic activities.
A typical solution for this first level support problem would be to have an employee portal with a static list of FAQs and a more or less powerful search incorporated into it. This approach works, three points for consideration though:
how good is our internal search?
is the portal available on mobile?
how do we assess the innovation potential of the above solution?
Enter the FAQ bot, with the following promises:
leveraging state of the art NLU for question / answer matching
availability on "any" channel (e.g. web / portal, your enterprise communicator)
ease of use by a non-technical business user (e.g. adding new questions, evaluating and tuning answers etc.)
Sounds interesting doesn't it.
How does the FAQ scenario differ from the transaction solution described in the first part? The main difference is that we use Microsoft QnA Maker which leverages Azure Cognitive Services in the background for the NLU tasks at hand. As you provide a set of questions and answers and the algorithms trains itself to be able to provide good answers to questions which can differ in terms of syntax (e.g . through spelling errors, different gramma etc.) but are same semantically.
Another difference is that in this scenario Microsoft provides a UI for creation of the knowledge base, the possibility to test the algorithm and an endpoint to consume the service.
In addition to the standard features available out of the box from Azure the RHR FAQ Bot integrates with your backend system to determine the knowledge base to be used for a given target population (e.g. our sales having slightly different bonus programmes then the dev team and hence different answers if similar questions) or the language of the user (e.g. answers in English or German). We offer as well support in creation and the tuning of the knowledge to cover your most burning scenarios.
With the features like out-of-the box NLU, availability on every channel, dynamically assigned QnAs, the QnA chatbot has the potential to dethrone more old-fashioned approaches to providing a scalable solution for employees' most frequent questions.
Interested in the look and feel? Check out this short demo to see how the bot performs.