Little butlers living in a box at the bottom-right side of the Moodle window. Teachers, but students too, can ask the ‘butlers’ to remind them of something for later on or to look up a resource or a piece of information on public or private repositories. More specialized bots can track performance, outcomes, and moods, or provide guidance for higher-level tasks. After a certain period of time-tracking and correlating variables, a bot could even predict and make recommendations to increase productivity and the quality of learning outcomes. On the commercial side of things, chatbots can speed up transactions, answer questions, and make cross-selling recommendations. A regular HAL 9000? That might come up along the way, but such a thing is not a requirement for getting started today.
This is, broadly speaking, the experience many of us are bound to enjoy in a not-too-distant future. While the idea of smart text or voice-based assistance has been out there for a while, the sprawling chatbot business has never been as active as it is today.
Can the momentum be carried into EdTech and Moodle? If the incentives are right and the conditions permit it, there is no reason why it cannot. Below, we propose a basic list of ingredients to make chatbots LMS-ready.
Is there something missing from our recipe? Be sure to let us know in the comments.
The use cases
By design, chatbots’ purpose is to give help and support to the user, usually in the context of a specific activity or user task. There are operations where chatbots happen to excel, but some tasks would benefit from more visual interfaces. Possible examples of best-case chatbot use are:
- Instant feedback and support: Frustrated users can quickly use a chatbot to vent their issue. The chatbot could offer some guidance and provide a basic walk-through. At the very least, it makes the user feel heard.
- Automating workflows or “macros”: The chatbot could track the user’s sequence of activities, and offer to take care of it when the user asks for it using a keyword. The user could design their own macros with help from the chatbot.
- Moderation: Added in group chats, it can track duties and responsibilities, record voting and decisions, and enhance teamwork in general.
The choice of a chatbot as an interface needs to be weighted against the existing alternatives. Giving a full lesson through chat, for example, might not be as engaging as through the many activities Moodle offers. Look at how chatbots were chosen and implemented in processes such as college admissions, performing health procedures, adherence to company policies, and recommendations and purchase processing.
The business models
Thinking about chatbots as a supportive technology keeps most of them away from concerns about financial sustainability. In Moodle, however, there are always at least a few developers looking for ways to monetize chatbots by themselves. We are looking forward to innovation in this area since, as we pointed out, evidence for self-sustaining plugin development is non-existent.
This issue is currently being explored by chatbot developers in platforms at large, and even in the broader app development field. The spectrum of answers today for self-sustaining chatbots offers the following possibilities:
- Donations and “tips”
Any of the previous options can also be included in a “Freemium” model. As we’ve highlighted before, freemium takes extreme care as to the signals it sends customers, especially if the person using it is not the person footing the bill.
Nevertheless, as often happens with new technologies that play a part in the interaction between people, it can be worthwhile to allow a large userbase to interact with the technology before developing uninformed revenue paths. Rough “bots” can be nurtured through constant user feedback, even to uncover income streams not previously conceived.
The development platforms
Part of the success of the modern wave of chatbots is the broader availability of platforms or development frameworks. As competition rises, they make increasing effort to make the process of building a functional bot simpler and friendlier.
The pressure to dominate the landscape is high for many platforms as they envision chatbots will be an important criteria for choosing a communications or messaging platform. While this can be beneficial for users and developers, it means that interesting features may not be immediately available within Moodle. The most prominent example of this is Facebook Messenger.
Fortunately, the ecosystem of messaging apps has allowed some openness. Today, several platforms allow you to build a chatbot which then can be brought into Messenger, Slack, Kik, Telegram, Skype, and more. (I’m under the impression that “Chatfuel” is where it’s at.)
When discussing chatbots, there is often the implicit assumption that the bots should be able to understand and interact fluently with humans. In reality, a chatbot could be useful even if it recognizes only a small set of commands. More sophisticated understanding capabilities are likely to become expected in ongoing generations, but at this point in time, the field is wide open.
All the above being said, a chatbot with too limited command recognition capabilities can be frustrating and could render the tool unusable. The recommendation going forward here is to build a chatbot incrementally, that over time can recognize more messages and consider more contextual cues, and on the other hand can respond as faithfully human as possible.
Being able to recognize and correctly interpret human inputs, particularly those stemming from what could be considered “conversational,” has long been the goal of the Natural Language Processing (NLP) field within Computer Science. But to carry out an ongoing interaction, there is the complementary field of Natural Language Generation (NLG). The NLP-NLG landscape varies, but it’s constantly improving.
In Moodle, it is technically possible to create a user with access to messages and forums. Using APIs, it could even hold conversations with other users. But to date, no application or tutorial is known to exist.
The community (aka, the humans)
Moodle was always envisioned as a teacher’s support system. No matter how relentlessly we pursue innovation, the role of the educator is not in doubt. When it comes to Moodle, questions like “When will chatbots replace teachers?” might not be the best use of our time.
The first Moodle chatbot would need a lot of support from the community, which could be as straightforward as interacting with it. Receiving and responding messages, along with feedback, is a slow-but-steady path to success.
On the forums, Moodler Inkar Alshimbayeva is looking to start a conversation about how the community can help open up the doors for chatbots in the LMS. He has three starting questions:
Do you think having chatbots would enhance online learning?
What problems of educators would it help to solve?
Would you want to have one as part of your online platform?
In the end, whether or not there is a future for chatbots in Moodle is largely a matter of how the “do-ocracy” that rules the community, and any new developments, responds to attempts by the entrepreneurial risk-takers within.
This Moodle Practice related post is made possible by: eThink Education, a Certified Moodle Partner that provides a fully-managed Moodle experience including implementation, integration, cloud-hosting, and management services. To learn more about eThink, click here.