CASE STUDY
Gargash Ai
AI-Enhanced Recruitment Workflow for Gargash
How our Ai powered recruitment workflow helped hiring managers at Gargash recruit and hire more efficiently.
Introduction
Gargash, a leading organization in its industry, sought to optimize their recruitment process to enhance efficiency and streamline operations. They aimed to leverage AI to automate various aspects of recruitment, from job description creation to candidate screening.
Challenges
The goal was to create an AI-driven recruitment workflow within Microsoft Power Platform, integrating GPT-4 to assist hiring managers in generating job descriptions, screening questions, and sorting candidates efficiently.
Inefficiencies in
recruitment processes
Too Much Manual
Labor Involved
Solution
We developed a comprehensive recruitment workflow within the Power Platform, embedding the GPT-4 model into Power Apps. The chatbot is linked to Langchain and the Pinecone vector database.
The recruitment workflow begins with the hiring manager opening a new position. Once approved, the hiring manager specifies the position’s requirements by filling out a form. Upon form submission, a chat session is initiated within the Power Platform, generating an initial draft of the job description. The hiring manager can continue to chat with the chatbot to refine the JD. Once finalized, the chatbot provides an approval button.
Upon approval, the chatbot generates screening questions based on the job description, designed to be close-ended for LinkedIn integration. The hiring manager can further refine these questions via the chatbot. After finalizing the questions, the hiring manager can use a button to trigger an automation that would post the job on LinkedIn.
After the job is posted, there is a waiting period for CV collection from LinkedIn. The collected CVs are then screened and sorted by the bot based on the job description and screening questions. The hiring manager can review the relevant CVs and shortlist candidates for interviews.
To streamline the interview scheduling process, the bot is integrated with Google calendar, checking availability and sending calendar invites directly to the candidates once they are shortlisted.
Every interaction with the chatbot involved backend prompt engineering to ensure the bot retained and applied the manager’s requirements consistently. For instance, when the hiring manager refined screening questions, prompt engineering rules continuously ensured those questions were appropriate for LinkedIn, maintaining the integrity and relevance of the prompts throughout the process. The initial form used to generate the job description is another prompt engineering layer that translated the hiring manager’s requirements into precise prompts for GPT-4.
Tech Stack Used: 
Your big ideas need a great team.
Book a free exploratory call with our experts today.
