OnboardingAI

Inspiration

Many companies do not have time to train new hires and walk them through every step. We want to build a continuously learning AI that solves this problem by learning from previous hires and answering basic questions before getting managers involved.

What it does

OnboardingAI helps with this process by learning from previous hires and answering basic questions based on guides/resources provided by the company. Resources and guides will be uploaded by the user, and whenever the new hire has questions, the AI will try to answer based on the resources it has. If the AI determines that the resources are not sufficient to answer the questions, it will send a ticket (via Intercom) to ping the managers for human help.

How we built it

We built the full stack using Lovable and used AGI Inc agent for our main AI. Users will be able to upload files from the frontend and basically talk (chatbot) to the AGI agent when they need help. The agent will determine whether its solution fails based on the frequency of repeated questions or if it cannot answer the questions. When the agent detects a failure to respond, it will automatically create a ticket via Intercom that sends a message to the managers stating the issue that the new hire is facing.

Challenges we ran into

The integration of the AGI API was difficult because we kept having timeouts and we also struggled with connecting our AI agent with the frontend.

Accomplishments that we're proud of

We were able to create a functioning frontend and backend; however, they are not connected.

What we learned

We learned how to code with unfamiliar tools in a limited amount of time and worked with continous AI for the first time.

What's next for OnboardingAI

We would like to connect the front and backend so we can interact with the AI using the chat feature in our frontend. We are also looking to incorporate it in Slack or Teams so the AI has more context from managers and previous hires.