Knowledge is power, as they say. But in any organization, there are a great many unknowns. Using FAQs or a knowledgebase is not always effective. Thankfully – as technical guru Vinay Joseph explains – Micro Focus IDOL has the answer.

Change means uncertainty; uncertainty needs knowledge

Over the past year, the corporate world has been turned on its head. More than ever before, organizations must provide support and clarity for a newly distributed workforce and provide better customer experiences through virtual channels. That means answering a plethora of questions as people adapt to new ways of working. But the key question is: what kinds of questions can an organization answer? Rarely will you see the same question asked the same way twice. So how does an organization automate such a complex task?

This is one of the advanced data analysis situations Micro Focus IDOL addresses. The IDOL virtual assistant uses Natural Language Question and Answering (NLQA) techniques to intelligently and automatically identify questions and match them to pre-existing answers delivering a customer-friendly knowledge base that enables faster resolution of a variety of queries.

The IDOL Answer Bank

IDOL has a component called the Answer Server that combines multiple features to create a seamless query response for users. In this post, I will be focusing on Answer Bank.

Humans are better than computers at answering questions with information that does not directly answer the question the way it is posed. IDOL Answer Bank is built to change that. Pre-established FAQs are added to a database. When queried, IDOL uses natural language processing to find the entry that best matches. These FAQs can be on a webpage, as rows in a database, or even as unstructured texts in a word document. To IDOL, the configuration doesn't matter as there are several approaches to indexing them using a variety of IDOL's connectors. The image below shows an FAQ hosted on a webpage.

Once the content has been ingested, IDOL converts each of these questions and their corresponding answers into pairs and stores them individually in the Answer Bank. See the image below of FAQs stored in the Answer Bank.

The above UI, which supports the ability to view these FAQs, is called the Data Admin UI. An administrator can manage these FAQs by editing them, adding new questions, or building out equivalent questions (see below).

Supporting Human Nature

A vital aspect of supporting freeform search queries is that the user decides what question to ask. And the nature of the language means there are several ways to ask the same thing. For example, the question, "How do chatbots work in IDOL?" could also arrive in the form, "How does IDOL handle conversational search?" The Answer Bank will be able to match these two questions with a medium-to-high likelihood as equivalent questions. This can be tailored by the administrator based on common search and response preferences to further improve the quality of interaction.

From the Find Open Source UI: the user asks a question just like they would make a typical search query, as seen below.

The answer to the question is then presented above the organic search results as can be seen in the image below.

Using IDOL to Improve the Answers

Since a question can be asked in a way that may not be matched to the original question in the Answer Bank, the Data Admin UI will pick up the unanswered question and allow the Data Administrator to map that unanswered question to an existing question in the answer bank. Additionally, IDOL can help suggest similar questions by conceptually mapping them. See the image below.

This is achieved by capturing the unanswered question in the incoming question queue, which has been designed to capture questions asked from “Find” or any UI which does not have an answer in the Answer Bank. See the image below of questions captured in the incoming question queue.

The feedback mechanism can help administrators capture questions asked from the UI which do not have answers and provide those answers. The upside of this approach is that search administrators communicate directly with their users and know the unanswered questions asked on the search application. The benefits to an organization are potentially significant, and include:

• Reduction in reliance on human support agents
• Improvement of response time, reducing downtime hours
• Uniformity in response quality

If you like to see the above workflow (from search to question curation and matching) in action, please see the following YouTube video where I am demonstrating it:

https://youtu.be/MHfc1nHxaxI

Learn More

• If you like to know more about IDOL, check out our IDOL product overview page.
• Learn about the latest IDOL 12.7 product release
• Take a look at this IDOL NLQA demo

I would love to hear your comments or questions. You can find me on LinkedIn and on Twitter at @josepheternity.