Friday, September 17, 2010

D2L Intelligent Agents

My first look at Intelligent Agents in Desire2Learn.

Having spent some time working on a Data Analytics project for MnSCU that would have served as an early alert system for online students not making good progress, I've been curious to see how deep the functionality runs in the D2L Intelligent Agents (IA) tool. The short answer is that there's some basic functionality there that might prove beneficial to some users, but the tool falls short of being a full featured Army soldier (you know, Be All That You Can Be!). (CC-by Flickr photo by Sarah G...)

I've recorded a narrated slide show that will illustrate the basics. In the slides you'll see: (a) where to access the IA tool, (b) how to create a new agent, (c) how to write a customized (sort of) email to be triggered by the agent, (d) how to schedule (or not) the agent to run when you want it to, (e) what happens when the agent is run, (f) and some pros and cons about the IA tool.

Click the green triangle Play button at the bottom of the frame to listen to the narration for the slides (or not). For better views of the screen shots, click the Full Screen icon in the lower right corner.

Basically, there are a few good things in this tool. You can receive notifications (and/or also send them to students via their D2L email address) for the lack of login activity for a specified number of days. You can also receive notifications when a student has performed poorly (or performed well, if you prefer) on a particular assessment that they have completed.

There are several shortcomings that I feel need to be pointed out. The IA tool could be very useful if these things are added in the near future:

1- The login agent only works at the system login level, not at the course access level. However, faculty will mainly want to know who isn’t accessing their specific course, which is where the agent was created in the first place, so this is only logical. With version 9, D2L added information to the User Progress tool to indicate the last time the student accessed the course homepage. Now they absolutely need to make this information available for the IA tool. Slide 14 specifically addresses this concern.

2- Currently the release conditions for an agent require an action on the part of the student, such as you took a quiz and got a high or low score, etc. However, it is usually more valuable for the faculty to know (and for students to get an email) when they are not doing something – didn’t take the quiz, haven’t posted to discussion forums, etc.

3- There isn't a way to create an agent that looks at overall user progress based on the overall grade book score, such as "you have only received 50% or less of the points available so far in this course, and your immediate attention is needed." The agent needs to be available to run on an overall view of user progress rather than just on one or more individual items in the course.

I contacted D2L about a future road map for the development of the IA tool. I received a reply from Matt T. He said, "I don’t have any details on specific time lines to address any of these or an overall road map for the tool, but I agree all three would be great additions."

My first grade for the IA tool is a solid C (I was always a tough grader, just ask my accounting students from back in the day). It is a step in the right direction, and has really great potential moving forward, BUT, it does need to move forward with additional enhancements to really meet expectations.

2 comments:

Dr. T said...

I absolutely agree that the inability to create an intelligent agent based on NOT doing something, like a quiz, is a MAJOR flaw. It seems to me that whoever created this feature did not think very carefully as to how an instructor would actually use this feature.

Theo said...

Agreed, a C is all it's worth at the moment. I look forward to seeing this tool mature. communications and prompts based on grades, student AND staff activity/inactivity.