How We Fixed 5,000+ Duplicate Time Entries in a QuickBooks API Integration
A misconfigured pagination parameter caused 5,000+ duplicate time entries in a QuickBooks automation. Here's how we diagnosed and fixed the STARTPOSITION bug.
Saving $32,000 / year to a Travel Agency Using AI
$32,000 in annual savings on AI software costs
60% savings on Management labor
Managers and agents can see their performance in real-time
The Client
This is a travel booking service that operates call centers to assist customers with flight bookings and travel arrangements. Key details about the client include:
- Call Volume: Averages 10 calls per minute during operating hours (6 AM EST to 10 PM PST).
- Employees: At least 100 employees per call center, with 4 call centers located in Latin America (Colombia), India, and the Philippines.
- Software: Ringba and Dialpad for call management, Zoho CRM for reporting and customer management.
- Partnerships: Contracts with airlines for preferential deals.
The Problem
They faced several challenges in managing their growing call center operations:
- Inefficient agent performance evaluation: unable to assess every call due to high volume.
- Lack of real-time feedback for agents: no immediate performance metrics available.
- Difficulty in identifying underperforming agents and top performers.
- High agent turnover rate.
- Inconsistent call quality and customer satisfaction.
- Scaling issues: inability to maintain management overhead as call volume increases.
- Cost management: Fixed costs for AI services during periods of lower call volume.
The Key Performance Indicators (KPIs) of focus were:
- Call completion rate.
- Customer satisfaction.
- Sales conversion rate.
- Average handle time.
- First call resolution.
The Solution
We implemented a comprehensive call analysis and reporting system that addressed multiple aspects of this Travel Agency's operations:
- Ringba and Zoho Integration. Developed a custom integration between Ringba and Zoho CRM to create tickets automatically when a call is answered.
- Overcame a technical challenge where Ringba API didn't provide the agent's identity upon call connection.
- Implemented a solution that rapidly queries Ringba’s API to determine which agent picked up the call, then creates and assigns a ticket in Zoho CRM with relevant customer information.
- AI-Powered Call Analysis:
- Implemented AssemblyAI for call transcription and analysis using their Lemur feature.
- Conducted a comprehensive cost-benefit analysis of different AI solutions:
- AssemblyAI: Offers both transcription and advanced call analysis (Lemur), but at a higher cost.
- Deepgram: Considered for potential cost savings on transcription.
- OpenAI/ChatGPT: Found to be most cost-effective for analysis but prohibitively expensive for transcription.
- Developed a hybrid approach to optimize costs:
- Considering moving call analysis to OpenAI for significant savings ($1,000/month).
- Retaining AssemblyAI for call transcription due to its cost-effectiveness compared to alternatives.
Because they now have transcription and intelligence (how well agents performed on calls, correctly answered customer questions, etc…) we could build:
- Agent Performance Scorecards:
- Developed real-time performance scorecards for agents, displayed both to the agents and their managers.
- Metrics include calls completed, hang-ups, customer satisfaction, sales conversions, and more.
- Advanced Reporting and Management Tools:
- Created customized reports in Zoho CRM for managers to assess individual and team performance.
- Implemented filters to quickly identify negative calls, underperforming agents, and top performers.
- Developed aggregate reports for call center performance over various time periods.
- Cost Optimization: we worked with the client to negotiate better contracts with AI providers to manage costs during low seasons.
- Negotiated an annual contract with Pipedream, resulting in significant savings on integration and automation costs.
- Continuously analyzing and optimizing AI usage to balance performance needs with cost considerations.
The Results
The implemented solutions led to significant improvements:
- Potential savings of $12,000 per year by optimizing speech-to-text services.
- Additional savings of $10,000 per year through contract negotiations with Pipedream.
- Reduced management expenses by at least 60% through improved oversight capabilities.
- Improved scalability of the workforce without increasing the management layer.
- Enhanced ability to assess agent performance and provide real-time feedback.
- Better customer service through real-time access to customer information.
Tech Stack
- Ringba: Phone system.
- Dialpad: Call/Agent management.
- AssemblyAI: Speech-to-text transcription and call analysis (Lemur feature).
- Zoho CRM: Reporting, ticketing system, and customer information management.
- OpenAI: Evaluated for potential use in call analysis.
- Pipedream: Workflow automation and integration.
Alternative Tech
- Deepgram considered for potential speech-to-text transcription.
- Claude.ai for call analysis.
Want something similar for your call center? We can help!
Let’s talk!
Made with Bullet