Case study
Call centre expanded to 2,500 employees
A customer service provider grew from 800 to 2,500 employees in 18 months without enlarging its recruiting team at the same rate.
Overview
ContactCenter Pro is a provider of outsourced customer service processes. After winning several major contracts, the workforce was to grow from 800 to 2,500 employees within 18 months. The existing telephone interviews tested communication skills and language proficiency but took a great deal of time. The company automated these first interviews with AI voice interviews, achieved 3x growth and at the same time improved the quality of hire and first-year employee retention.
Results
- Time saved 82%
- On screening activities
- Employees hired 2,500+
- In 18 months
- Candidate rating 4.8/5
- Interview experience
The challenge
To grow from 800 to 2,500 employees in 18 months, ContactCenter Pro had to hire around 95 new employees every month. That required 400+ telephone interviews per month. Even when concentrating entirely on these calls, the 10-person recruiting team could handle only about 250 a month. In a call centre, communication skills, language proficiency, tone of voice and spontaneous problem-solving have to be tested in conversation; CVs say little about them. The existing 30-minute telephone interviews captured these qualities but were time-consuming and hard to standardise. Some recruiters assessed very generously, others very strictly. These differences led to inconsistent hires and a first-year attrition rate of 45%. That caused repeated hiring and training costs and disrupted operations.
- High hiring demand: 95 new agents a month required 400+ telephone interviews against a capacity of 250
- Communication as the core criterion: success in a call centre requires assessing voice, tone, articulation and conversational ability
- Inconsistent screening: recruiters applied different standards, which led to 45% attrition in the first year
- Several languages: agents fluent in German, English, French, Italian and Spanish were needed for different client contracts
The solution
ContactCenter Pro used Talky to standardise and automate the first telephone interviews for all call centre roles. The interviews captured clarity of communication, active listening, problem-solving under pressure, empathy and language proficiency. They lasted 15-18 minutes and followed the structure of a customer call. They included several practical cases, for example: 'A customer calls because they have received the wrong item and is very frustrated. How would you help them?' The AI assessed content, tone of voice, speaking pace, clarity, empathy expressed through the voice and the approach to the solution. For multilingual roles it switched language to test proficiency during the conversation. Applicants with good communication skills and an aptitude for customer service were automatically given an appointment for the final interview with the hiring manager.
- Assessment based on conversation scenarios: realistic customer service cases test problem-solving and communication under pressure
- Voice quality analysis: the AI assessed tone, pace, clarity and empathy, not just the content of the answers
- Multilingual assessment: language switching during the conversation to test proficiency in German, English, French, Italian and Spanish
- Standardised scoring: the same criteria for all candidates, regardless of the individual recruiter
How the rollout went
New client contracts were due to start in 90 days, so ContactCenter Pro had to expand its recruiting at short notice. The VP of Operations, the head of recruiting and successful team leaders supported the rollout. They developed practical interview questions and analysed call recordings of successful employees for this purpose. They looked for communication patterns, approaches to solutions and vocal characteristics associated with good job performance. These observations fed into the conversation scenarios and assessment criteria.
Capturing requirements (days 1-5)
Analysis of call recordings of employees from the top and bottom performance quartiles to identify distinguishing communication patterns. Interviews with operations managers and team leaders to define the ideal agent profile. Documentation of key competencies and creation of a scoring scale.
Conversation scenarios and calibration of the voice analysis (days 6-12)
Development of 8 realistic customer service scenarios covering common types of problem. Testing of the scenarios with 30 current agents to calibrate the AI scoring. Fine-tuning of the voice analysis so that it distinguishes nervousness from genuine communication problems.
Connecting and testing systems (days 13-18)
Integration with the Workday ATS. Set-up of automated interview invitations. Set-up of the multilingual interviews and testing with native speakers of each language. Definition of scoring thresholds based on performance data from current agents.
Pilot and rollout (days 19-25)
Trial with 100 applicants, run in parallel with traditional screening. Comparison of the AI assessments with the recruiters' assessments. Final adjustments to the scoring. Use across the entire selection process from day 26.
The results
Interview numbers, time to hire and employee retention changed markedly. Over 18 months ContactCenter Pro interviewed 8,437 candidates by AI interview and hired 2,547 agents, exceeding its growth target. The time from application to completed screening fell from 8 days to 3 hours. That sped up selection considerably and reduced the number of candidates dropping out. The quality of hire improved markedly. First-year employee retention rose from 55% to 75%, so fewer people had to be rehired and retrained. Operations managers reported that agents who had passed the AI interview were consistently well prepared for the actual demands of call centre work. Standardised scoring removed the differences between recruiters that had previously led to fluctuating quality. Recruiters no longer had to conduct 400+ telephone interviews a month and could concentrate on looking after applicants personally. That improved the applicants' experience of the selection process and the employer brand.
“Talky's AI interviews actually do a better job of assessing communication skills than human phone screens, because they are completely consistent and unbiased. Every candidate gets the same scenarios and is evaluated against the same criteria. It is not just faster - it is more accurate. Our agent quality and retention have never been better.”
Marcus Schmidt
VP of Operations, ContactCenter Pro
Key takeaways
AI interviews with conversation scenarios are suitable for assessing communication skills and aptitude for customer service
Voice quality analysis (tone, pace, empathy) can be automated and provides a consistent assessment of all candidates
Standardised AI screening removes differences between recruiters and improves quality of hire and fairness
Multilingual AI interviews enable BPO providers to grow efficiently while serving different client requirements
The higher quality of hire (75% versus 55% employee retention) delivers a greater ROI during growth than the sheer number of hires
See Talky with your jobs?
30 minutes, with an example from your recruiting.