ID | Requirements | Priority | State | Contributors |
---|---|---|---|---|
1 | Send requested resources to users for learning | Must | ✔ | Ryan |
2 | Understanding and responding to simple sentences | Must | ✔ | Ahmed, Luming |
3 | Pro-actively find time in their calender to complete the on-boarding | Should | ✔ | Ahmed |
4 | Set up session in the user calendar | Should | ✔ | Ahmed |
5 | Saving the user's on-boarding progress | Should | ✔ | Ahmed |
6 | Have authenication | Should | ✔ | Ahmed |
7 | Engage the user to on-board | Should | ✔ | Luming |
8 | Send proactive notifications to the user | Should | ✔ | Luming |
9 | Tracking user progress | Should | ✔ | Ryan |
10 | Using Power BI to good understanding of how many people are at each stage of the on-boarding and how many are complete | Should | ✔ | Ryan |
11 | Tracking user behavior | Could | ❌ | Ahmed |
12 | Allow the user to request tutorials directly | Could | ✔ | Luming |
13 | Have buttons for easier dialogues | Could | ✔ | Ryan |
14 | Mark finished tutorials | Could | ✔ | Ahmed |
15 | Having natural language understanding | Could | ✔ | Luming |
16 | Provide links to documentations for further details | Could | ❌ | All |
Following our design principles on HCI section, we have prioritized simplicity, intelligence and continouity.
The bot is easy to use, it is has a plain design, simply displaying the chatboxes as the dialogues ensue. To keep things clear and simple, we have made sure to not include any unnecessary dialogues for on-boarding in the bot. If the user would like to see more information about Office 365, a link is provided so they could explore it themselves as needed. This would avoid the bot overloading users with information that is not essential for the on-boarding, while still giving them the option to explore it themslves from the dialogues.
By utilizing QnAbot, our bot can response to simple user inputs directly as long as the the bot is opened. For example, the user can request a tutorial directly. This makes the on-boarding process more natural as the user can do things directly as he wishes if he is familar with the bot.
The bot remembers the user's progress and checkmark the tutorial if the user has already gone through it. The user would be able to see what tutorials he has gone through as he chooses what to learn next, allowing him to choose a unwatched tutorials without needing to check his progress elsewhere. Furthermore, if the user closed the bot in the middle of a tutorial, the bot will ask him if he wishes to continue the tutorial and return to it if needed, saving the user's time.
These design choices and features of our bot allow for an intuitive and simple user expereince that is also effective in teaching the user.
Functions for the bot have been implemented, and works as expected. Power BI also works well and naturally with Azure Application Insights to track users' progress.
From the testing we have done, Personal On-boarding Coach operates properly under various environments without crashing almost all the time. Instabilty of Azure can cause the bot unable to run, this happened a few times during our development.
Our bot is an online service built on Azure. Different platforms can access it through connectors. It should support all major platforms including phones as long as an access token is obtained for the platform. It also can be connect to any channel such as Facebook, Teams, Slack and so on.
With the tools we have available, our team has managed the project well throughout its progress. Azure Online Service allows us to set up more online functions like logging and test our bots easily. GitHub allowed us to code and work together smoothly. To keep communication between members consistent and strong, we met the client frequently to discuss the development of bot and made sure each of us had different work for the project every week. Although we have had several unexpected delays throughout the development due to Azure's being down sometimes and subscription running out, we were able to plan our time well and get ourselves back on track.!!
We are glad to have been able to work with our client Addin 365 for the year. The scope of our project gave us the oppotunity to work with new technolgies in bot development. With the support and help from our TA and client, we were able to work better to achieve the targets we had at the beginning of the project and have gained valuable software development in the process.
By utilizing MSgraph, the bot should be able to check users’ usage of different Office 365 tools like Teams and OneDrive. By tracking the usages of the tools, the bot would know if the user is currently using the tools or not. If not, the bot should ask the user to do so and ask him if he wants to learn that part again if he has forgotten about it. This can ensure users on-board as expected and do not go back to their old ways of working.
We wished to add links to documentations for each tutorials to help users with the on-boarding process, but were unable to due to limit in time.