By Erica Dumore
My name is Erica Dumore and I am a second year law student at Suffolk University Law School. Over the last five weeks I have been introduced to the world of coding, written my first line of code, and grappled with creating a bot, which you can find here. To write my bot, I used a tool called QnA Markup that allowed me to write code, edit it, and regenerate it into the bot linked above. Essentially, this tool allowed me to create an interactive decision tree by nesting questions and answers that are triggered by a users interactive responses. A copy of my written code is linked here.
As a beginner, a lot of refinement went into generating my final product. I was constantly editing and tweaking my bot to make it more fluid and user friendly. From start to finish, over the five weeks of work, I could not even attempt to count how many times I edited or altered my bot, I think the progress speaks for itself. Compared to my final bot, I came a long way. Below please find six copies of drafts should you want to see my progress throughout the five week period.
First Code Draft: Simply experimented with using Q, A, and X syntaxs.
Second Code Draft: Expanded the first draft to further utilize the GOTO syntax feature, placed javascript placeholders in potential places, and also added in a header and title.
Third Code Draft: Expanded the header and title, implemented the javascript syntax, and altered the end results.
Fourth Code Draft: Edited to ensure every possible outcome had a final response, added in a few website links to direct user, and put placeholders where the statutes and forms would be placed.
Fifth Code Draft: Implimented a mail merge document of the Fee Waiver Form so that the QnA code would generate answers from the code into the document.
Sixth Code Draft: Implimented all missing links, checked to verify they were active and edited for any mistakes.
A bot is only as good as its creator allows it to be. Throughout this process I quickly learned how sensitive coding can be, one missing colon or misplaced letter disables the whole bot from that point forward. I was very careful to repeatedly update my QnA to ensure it was properly generating and producing the output I desired.
With that said, I also conducted peer and attorney reviews, which were extremely helpful for integrating improvements. As with any task, it is very easy to believe you have an effective working project when you are in fact just too close to see a hiccup or a more beneficial way to organize/better present the information.
Peer Review:
This was the case with my bot, my fellow classmate provided me with feedback after testing my bot and explained that at that point in my drafting stage, without realizing it, I had directed all of my pathways to the two same generic result responses, which as you can see in my final bot is not the result I had been seeking.
Ultimately, my reviewer helped me realize that my two generic results could be broken down further and made simplier, allowing the user to better understand their results. This feedback truly helped me to generate a more accurate set of results that the users could understand more thoroughly.
Attached, you can find a copy of my first peer review email correspondence, which lead to my second draft bot integrating the feedback.
DC was CC'd on these email chains between September 10, 2017 and September 16, 2017.
Attorney Review:
Additionally, in our ever advancing society, technology is the resource that most people turn to when they face an unanswered question. My goal of this class project was to generate a model user-friendly bot that could help a target group of people in New Hampshire who are seeking legal answers. Therefore, who better to ask to test my bot than attorney's themselves.
I reached out to attorneys who practice in the State of New Hampshire to ask them if they would be willing to review my bot and provide me with feedback, positive and/or negative, to help make the bot more user-friendly, more relatable to the target audiance, and/or more effective overall. I received responses agreeing to review the bot upon completion and provide me with feedback. At this point, I am waiting to receive said feedback.
This email chain was forwarded to DC on September 26, 2017.
Finally, although a 'final' version of the bot is attached above, there are still steps that I would like to take to polish the bot before I feel it would be real-world viable. Below is a list of changes I believe would need to be addressed:
(1) As mentioned above under User Testing, I have reached out to attorney's in New Hampshire and received positive responses and willingness to provide me with thoughts and potential improvements, however, at this time I have not received attorney's feedback on the bot and therefore, before making my bot real-world viable I would like to implement my final version to include the any attorney feedback.
(2) Additionally, at this point my bot generates a fee waiver form, auto-filling only four answers. Before making my bot real-world viable, I would like to expand this to auto-fill the complete form and I would like to have my bot generate the appointment of counsel form when involved in a criminal matter.
Therefore, although my bot is functional and able to provide users with proper feedback, I am aware that my bot has room for improvement before it is real-world viable and would take the necessary steps mentioned above before promoting it to users.