Inspiration
For HR companies, the most tedious part is to screen resumes and do pre-phone interviews with a lot of potential candidates. This not only costs time, money, and human resources but also makes HR representatives feel exhausted, since they have to talk on the phone for long periods of time.
Because of the above issues, we would like to introduce DashaHR. DashaHR is the process that uses Dasha's Conversational AI to pre-screen potential candidates to check if they are qualified for the next round in person interviews with managers.
What it does
Using DashaHR, users are interacting with AI that has human-like voice and natural talking flow. The AI asks potential candidates questions about their skills and education. Based on the information, the AI can determine if the candidates are qualified for the next round of interview or not. Afterwards, the AI will set up in-person interviews with the qualified candidates.
How we built it
DashaHR was built using the conversational AI of Dasha. Thanks to Dasha, the conversations between humans and AI are very natural. We create intents, data, and entities to detect the leads as well as the patterns in humans' speech. The Dasha Script Language (DSL) also enables us to build dialogs and transitions in the conversations smoothly.
The conversation is structured orderly with certain conditions being triggered in different nodes and detected by intents. The conversation is transitioned to different routes based on the intents of the humans. AI can sort if the humans meet the requirements to go a different route of conversation (setting up next round of interviews or not).
We need to create different intentions and macros to train the AI. The most difficult part is to have a good source of phrases or intents to predict how the humans talk or discuss in conversation.
Challenges we ran into
The Challenge we had was to know and learn Dasha for the first time. We had to go through different documentations to understand the flow and how to develop the conversations. Further, the time constraint for the hackathon is also a factor that makes us feel worried that we cannot finish the project on time. We also had multiple error issues and we had to debug to resolve each error one by one.
Accomplishments that we're proud of
We were able to resolve all the errors that we encountered. We also learnt about Dasha and how well the conversation that Dasha provided. We were able to make the whole process work, creating the demo, setting up blog, and video.
What we learned
We learnt about different functions of Dasha AI and how to build different scenarios that can be used together with Dasha. We discovered that using Dasha the AI conversation flow is very natural and makes me feel like I am talking to the real person. The voice is also natural and gives users confidence to talk with the AI.
What's next for DashaHR
We would like to incorporate the process in automation process in which machine gets info of candidates automatically and then calls qualified candidates if their skills match the requirement of the job.
Log in or sign up for Devpost to join the conversation.