DIGITAL TRANSFORMATION TODAY
Bringing Microsoft Chatbots and Other Virtual Agents into Office 365
May 18, 2021
Microsoft Chatbots are quickly becoming the new norm within Office 365 apps such as Teams and SharePoint.
The first part of this series talks elaborately about understanding Bot design language and processes. Click here to read part one.
Before diving into mechanisms to bring the chatbot into your ecosystem, here are some common steps you will want to consider when bringing chatbots into your organization or team:
- Gather your use case. What problem are you trying to solve? An everyday use case would be: as an organization, we need to have a 24/7 resource available to answer company policies such as COVID 19 vaccinations or benefits.
- Consult with an expert. Once you have your use case established, reach out to Withum to go through your pain points and provide recommendations and best practices.
- Information and Language design/architecture. Identify the type of chatbot as well as language flow and branching (see Series 1).
- Platform selections (current blog). Depending on your needs, you will have to decide whether to use a third-party tool or not to bring the chatbot in Microsoft, as well as whether to display the chatbot in SharePoint and/or Teams.
- Bot selection. Various solutions are available, a robust and easily customizable Power Virtual Agent (will require a technical resource), quick and easy bot setup with QnA Maker, sophisticated yet easy setup with third party tool like AtBot (will require a few hours of a technical resource).
- Communication plan. As part of change management, staff needs to know why they will use the chatbot, how it impacts their work and the long-term vision that can encompass early on through a communication plan and execution.
- Implement (current blog). Depending on your platform selections, there are different mechanisms to deploy the chatbot.
- Test/Training. Within the Microsoft QnA Maker or in third-party vendor tools such as AtBOT, you can test and train the bot to refine your questions and answers. You can also share the chatbot with a pilot group of users and get their feedback.
- Once you have received feedback during your testing, it’s time to launch. Make sure to have a communication plan ready ahead of time.
- There are mechanisms in receiving feedback, either through surveys or analytics, to auto-gather chatbot user questions, including unanswered questions.
- Keep the chatbot alive!
Now let’s talk about tips and tricks in deploying the chatbot.
- AtBot (link)
- Recommended use case
- A chatbot that goes above and beyond the out of the box features such as kick off chatbot with a workflow or an adaptive card
- Automate your business cases and connect to your services
- Integration with QnA Maker, LUIS and Power Automate
- Easily package solution and deploy in Microsoft Teams
- Deploying in Microsoft SharePoint Online
- Deploying in Microsoft Teams
- AtBot makes it very easy to deploy to Teams. The AtBot admin portal creates a ready-made installation package that you can download into Microsoft Teams.
- Who can do this? A Teams Administrator and AtBot administrator.
- Microsoft QnA
- Recommended use case
- A knowledge base (think of replacement of excel UI) that lists out all my questions and answers.
- Question and Answer bank tied into AtBot or in your chatbot.
- Deploying in Microsoft SharePoint Online or Microsoft Teams
- QnA Maker services needs to be created in Azure.
- Who can do this? An Azure or resource group Administrator.
- Withum’s Chatbot for SharePoint
Technology and Digital Transformation Services
- Decide on a site the chatbot should exist