Table of Contents
Introduction
You’ve finished setting up your Departmental Assistant. Now you’re probably wondering: what’s next?
This guide will walk you through the Admin Console, the web interface for managing UMD Virtual Agents. There are two types: Departmental Assistant and Virtual Study Assistant, both use the same Admin Console but with different emphasis. As an administrator, you’ll review the questions users ask the Departmental Assistant, along with its responses, and update or clarify content when needed to ensure the chatbot provides accurate and helpful information to visitors.
The Admin Console has many features—too many for one guide! We’ll cover the key ones relevant to Departmental Assistant, leaving advanced options for a separate guide.
To begin, visit your Admin Console and log in with your university credentials. Once logged in, you’ll see a list of courses or agents where you’re an administrator. If you have trouble accessing your Admin Console, just reach out to the DIT AI Solutions Team at dit-ais@umd.edu.
Overview of the User Interface
Here’s what you’ll see when you visit the Admin Console. Your course's Virtual Agent (Departmental Assistant) name appears in the top left corner next to the UMD logo. Additional options are available by selecting the yellow circle with the first letter of your first name, which is located in the top right corner.
Exploring the Navigation Sidebar
Question Review
The first section is Data Analysis. Under it, you'll find Question Review.
Once your Departmental Assistant is up and running and visitors have started using the chatbot, this is where you can review the questions they’ve asked.
Let's look at the user interface.
In the top-left corner, you’ll see a date selector. Choosing a date will display all the questions asked during that year and month.
At the top, you’ll see several tabs like ALL, UNKNOWN, KNOWN, and more. For now, keep the ALL tab selected.
Below that are the columns: a checkbox (☐), Actions, Timestamp, Intent Used, Question, Response, and Session ID.
Pay attention to the following columns.
- Timestamp: When the question was asked
- Question: The question asked
- Response: The Study Assistant's reply
- Session ID: Identifies a specific session (helpful for seeing if a user asked multiple questions during one session). In the example below, the same person asked the first two questions, confirmed by the duplicate session ID found in the Session ID column of the questions.
If you see multiple questions with the same Session ID (circled above), they all came from the same person.
Regularly reviewing chatbot responses can help you spot where visitors may be receiving inaccurate or incomplete information. When that happens, you can update your website content to help the chatbot deliver clearer, more accurate answers. This leads to a better experience for your visitors, with more helpful and reliable responses.
Data Sources
Next, underneath the Data Sources section, you'll see Scripted Questions.
Departmental Assistant typically gathers information by web scraping a list of URLs and their child pages that you provide. However, if the scraped website content doesn’t provide clear or complete answers for certain questions, you can supplement it manually. Use the Admin Console to add common questions and provide precise responses, helping the chatbot deliver better answers in those cases.
This is the user interface that you see when you click on Scripted Questions.
Initially, this page will be blank with no questions nor answers. To add a new question and answer, click the plus button in the top-right corner. Then, enter a question and its corresponding answer. You can add as many as you need. Under Actions, you see a large dash button. To delete a specific question and answer, select the dash button.
Analytics
The next section is Analytics, in which you will find the Virtual Agent Analytics subsection. This subsection provides insights, including the number of questions asked, unknown questions, thumbs up, down votes, and more. We’ll cover this in more detail later.
If your Departmental Assistant was only published recently, you might not see much data yet—but that will change as more visitors start using it.
Virtual Agent Config
The next section is Virtual Agent Config. It contains two subsections: Virtual Agent Config and Ingested Sources.
You may wonder why this is labeled Virtual Agent Config instead of Departmental Assistant Config. That’s because the Admin Console is used to administer all UMD Virtual Agents, including Virtual Study Assistant and Departmental Assistant, so the broader term is used.
Virtual Agent Config
This is one of the more useful sections in the Admin Console.
The first section includes a link to your Virtual Agent’s user interface. Below that, under Virtual Agent Authorized Users, you’ll see a list of all administrators for your Departmental Assistant.
Changing the UMD Virtual Agent's Name
When you open your UMD Virtual Agent URL, it launches in a new window, and the title bar will look like this:
The title bar displays a generic name, UMD Virtual Agent, and the chatbot shows a default greeting message.
Let’s start by customizing it to something that better reflects your department or audience.
First, scroll down to the System Configuration section at the bottom of the page. You’ll see the following:
The Agent Name field controls what appears in the chatbot's title bar. When this field is left blank, it defaults to UMD Virtual Agent.
In this example, we’ve customized the Agent Name to UMD Workday Virtual Agent and added a personalized greeting message.
Here’s how the UMD Virtual Agent looks now, with the custom name and greeting in place—much more personalized than the default version.
Understanding the Web Scraper Connector
Scroll back up to find the section labeled Connectors.
A connector links your Departmental Assistant—a generative AI chatbot—to external platforms.
For the Departmental Assistant, the Web Scraper is the primary connector. It also supports a Google Drive connector, but we won’t cover that here—we’ll save it for a separate guide.
You might be wondering, what exactly is a web scraper?
A web scraper automatically collects content from a list of web pages and their subpages, pulling in text that can be used by the Departmental Assistant. In a Retrieval-Augmented Generation (RAG) system, this scraped content is indexed so the assistant can search it and use relevant information to generate accurate answers. This allows the chatbot to stay up to date with your website content without requiring manual input.
The section below provides key details and usage tips for the web scraper connector.
Underneath, you can add URLs for the web scraper to scrape. The web scraper will visit child pages so you do not need to explicitly list out all the pages of your departmental website. There is a limit of 200 web pages that it will scrape. To increase that, please reach out the DIT AI Solutions Team.
To add a new parent URL to be scraped, click the red Add button in the top-right corner. Use the red circular arrow to refresh the list, and click the trash can icon to delete a URL from being scraped.
In the Sources column, you’ll see the number of pages scraped from the URL and its subpages.
Ingested Sources
The process of storing scraped information so it can be used by the Departmental Assistant is called ingestion. Once the web scraper completes scraping, you can see what it ingested by selecting Ingested Sources.
You’ll see an introductory explanation followed by three tabs: Ingested Sources, Broken Webpages, and Ingest Details. For now, just focus on Ingested Sources.
The following table lists all the sources ingested. The Type column is explained above. The Ingested column indicates the last time the resource was ingested.
Under the Included column, you’ll see a slider—but note the lock icon next to it. This means you can’t change the inclusion status for that source. While other connectors (like the Canvas connector) allow you to exclude individual sources from ingestion, the web scraper doesn’t support that level of selectivity.
If your website has many subpages, the list of ingested sources may be quite long.
Support
We've reached the final section of the Admin Console overview. If you have questions or need assistance, you can contact the DIT AI Solutions Team directly through the console, or email us at dit-ais@umd.edu.
Additional Resources
We’ve only covered the essentials of the Admin Console to help you get started with your Departmental Assistant.
To explore more advanced features and settings, check out the Admin Console Intermediate Guide. There, you'll find information on additional connectors you might find useful —such as the Google Drive connector.
If you need additional help or have any question, reach out to the DIT AI Solutions Team at dit-ais@umd.edu. You can also contact us if you need assistance in embedding a Departmental Assistant chatbot on your department’s website.