Achieved Key Result
93%
faster chat resolution time
Explore how ProjectFitter implemented CoSupport AI within Freshdesk and Freshchat to automate support workflows, improve response quality, and deliver faster, more consistent customer assistance across chat and email.
Achieved Key Result
93%
faster chat resolution time
Country:
USA
Industry:
Software Development
Integration time:
15 days
Support Team:
8 agents
Support Setup:
Freshdesk & Freshchat


The Client’s Profile
ProjectFitter is an AI-driven hiring platform that matches tech candidates with jobs, reducing time-to-hire by 40%. The platform offers advanced filtering, candidate scoring, and success-rate predictions for employers, while candidates benefit from personalized recommendations and cultural-fit scoring.
🤔 Challenge
ProjectFitter initially attempted to build its own AI support model using OpenAI’s API. While the prototype worked, integrating and maintaining the solution inside Freshdesk and Freshchat proved too complex and resource-intensive. With support volume increasing across chat and email, the team needed a fully integrated, scalable AI system capable of handling both real-time inquiries and email support without heavy engineering effort.
🏆 Goal
The primary goals were to reduce the average resolution time from five hours to one hour, automate 50% of all user inquiries, and seamlessly integrate AI into the existing Freshdesk and Freshchat environments.
💡 Solution
The team adopted CoSupport AI’s integrated support solution:
AI Agent — for structured internal documentation and consistent answer generation based on the centralized Platform Guide.
AI Assistant — for Freshdesk and Freshchat, handling real-time chat conversations and email requests with AI-generated responses, triage, and escalation support.
🔗 Integration
Freshdesk & Freshchat Integration
CoSupport AI is integrated directly with ProjectFitter’s Freshdesk and Freshchat environments, enabling the AI agent to resolve common questions, summarize conversations, classify issues, and escalate complex tickets as needed. Initially, ProjectFitter attempted to train the AI on historical tickets; however, outdated and inconsistent data led to inaccuracies in the training process. To address this, the CoSupport team introduced a structured approach by creating a centralized Platform Guide that included accurate procedures, feature descriptions, and best practices. Once this guide was connected to CoSupport AI, it became a reliable source of truth, producing consistent, high-quality responses across all support channels.
🚀 Achieved Key Results

Client’s Feedback
“CoSupport AI streamlined our operations with advanced customer service AI tools, automating about 70% of support tickets and shortening resolution time from hours to minutes.”

Yaroslav Burgman
Project Manager