You decided to create a page on social media to have better exposure. It gets noticed by more and more customers until you have hundreds of them! (hooray 🎉) But there’s a problem you didn’t anticipate…you now spend more than 30% of your time answering their queries rather than focusing on your actual business. Hiring a social media manager might cost you more money than what you currently earn so you also cancel that out.
Now, what if someone told you an AI could help you in situations like these? An AI which could answer customer concerns so you can go back to focusing on your passion instead.
Let’s say you start to look for FAQ system solutions online but you’re bombarded with all these different options. Traditional ones have been around for a while but most existing solutions rely on having the exact same copy from the sample questions you’ve fed them and comparing them to a person’s query. This leads to numerous unanswered questions because they are not the exact match.
For example, you added in your FAQ list “address please?” and the customer asks, “what is your address?”. They essentially mean the same thing but since the two sentences don’t match word per word, traditional FAQ systems won’t understand it and might fail to catch them.
What an AI model does is understand the query’s context and compare them with the list of questions you have. It doesn’t matter if it doesn’t match all the words in the sentence as long as they have the same meaning. You only need to provide several other variations of the same question to help the AI learn better. The more data you feed, the smarter it gets. This also increases the chance for it to answer most if not all the questions you ask it.
The Q&A API powered by aiah.ai gives you access to the latest AI technology making sure that your FAQs will be answered. This service is accessible through this link: https://aiah.ai/service?id=dcd56381-f352-4819-b5b9-0ff4e06cf359. Just create an account and you’ll have access to all available AI services including this one!
Here are the 3 quick and easy steps you can follow when using the AI Marketplace Playground:
Aside from the kb itself, also think of a collection name which is easy to remember since we’re using this later as an identifier before we can ask questions. Now go to the Question and Answer API “upload knowledgebase” endpoint and enter the following information. From the example below, we used “basic_info” as the collection name. Then, copy and paste this kb in the “Request Body” section and hit “test endpoint” at the bottom. It should look something like this.
Once the upload is successful, we can now use the “query qna knowledgebase” endpoint to test it out. Maybe we can try asking “what is your address?” and see the results. If you try it yourself, you should get a result similar to this one.
If you want to add this functionality into your own app or maybe your own chatbot, you can use the following snippet to upload your kb in your AI marketplace account’s project. This is a required step so you can send requests to the query endpoint.
The aiah key and client project id are provided details once you’ve signed up for an account. From this snippet, you can add more training phrases and answers as much as you want for the knowledge base. Later on you can still update this using the “update knowledgebase” endpoint so no worries if you can’t exhaust all the questions you want to add in the list.
Now all you have to do is to send a request to the query endpoint each time a user asks a question. Just retrieve the answer from the top result and you’re now good to go!
The results you get from the decoded data should be in this format