This is the story of my simple little Chat Bot Project, something that I wanted to turn into a big project.
Back when I was in college, we as students struggled to get the necessary documents, we were always trying to find exams from previous years from other universities and books or courses that better explain a subject, the problem is there weren't a go to place for this, you would spend hours looking in Facebook or some website and most of the time, we would come up empty. the documents were there, the idea was to make a chat bot that you can give it the documents you have, tell some basic info on what these documents are and it stores it so other person can find it easily, so the bot will handle hundreds of student sending and receiving.
I didn't have the courage to launch it as I thought a database like that would cost me money plus the possibility that the bot might crash and I was the only one who knows what's going on.
That's the story in short, of course it didn't stop there, as I thought the idea of a chat bot could be applied in various other domains. I will talk about that after I explain a little bit about the bot
My Chat bot was mainly to be used in Facebook Messenger where you ask it to find a document for you in French or English and it uses NLP behind the scene to extract key words that will later help to find the files in the database
For the NLP (Natural Language Processing), I used Dialogflow
I realized the same idea can be applied in many other ways, one idea that came in mind was to find Professional workers nearby, every worker (Plumber, Electrician, Painter...) would register in the bot with very simple process.
and then people would ask the chat bot to find them a Professional worker and it would search the Database and provide them with a list of worker and their phone numbers.
Now, this idea might not be very useful in other places, but in Algeria, there is really no better mean to find workers, the only other way is to ask around which is time consuming to say the least.
This chat bot app can even grow to include a rating system so that Good Workers are encouraged and displayed to users more often.
We would add training phrases in Dialogflow to ensure good NLP and keyword extraction, I created Entities for each Worker with all possible wordings in 3 languages to increase the chances of understanding what the user is looking for.
The image below shows that the Chat bot understood the words that we provided and started searching, no result is displayed yet because the Database that will contain all workers and their phone numbers was still in early stages.
One more idea that I thought about shortly after the previous one was a Job application processing app, where the chat bot will receive resumes/CV from all sorts of job applications and will chat with users to identify the type of job they are applying for (so this means the bot can handle a large amount of job selection processes and can reduce the time spent to produce a shortlist of candidates. companies who are hiring can provide us with a set of questions to be asked to users applying for the job, this way the bot will select possible candidates based on the answers of the users.
The idea to use a chat bot to process job applications is not new but I wanted to make custom made one for Algerians.
The video below shows an example of a Fake Hiring process where the Bot is chatting with user for Job application for Coca cola
Did I mention that the Chat bot can even send MEMES!
Thanks to Dialogflow, the same chat bot can be integrated with many other platforms
,This is slightly different than what I showed above, but I used the same principle of NLP with Dialogflow to create a voice assistant with the help of Actions on Google.
The assistant will find doctors and clinics for users for a specific location that they choose.
Again, this is made using a small database of doctors I scraped from website for Algeria.