How Deutsche Telekom Runs 95% More Consistent Technical Interviews — At Scale
"Half my interviews weren't getting detailed, structured feedback from panels. JobTwine Copilot changed that. Now every interview is captured, transcribed, and tells me exactly what went well and what didn't."
Anurag Tiwari
TA Head, Deutsche Telekom
95%
Improvement in interview consistency
99%
interview satisfaction score across panels
Ready to hit similar milestones?
At Deutsche Telekom today, the moment an interview ends, a structured feedback report is already waiting — scored, timestamped, and ready to act on. They can now hire fast and achieve consistent quality.
But it wasn't always like this.
Challenge
Deutsche Telekom Digital Lab was hiring at scale. The volume was there. The urgency was there. What wasn't there was consistency.
Every panellist came in with their own approach — their own questions, their own informal scoring, their own instincts about what a good answer looked like. Some rounds were rigorous. Others were conversations that felt productive but produced nothing useful afterward. There was no shared structure tying it all together.
The harder problem was what happened after the interview ended. Panellists were expected to simultaneously lead a technical conversation, evaluate responses in real time, take notes, and submit structured feedback — all within the same hour. Most couldn't do all four well. Feedback came in late, incomplete, or not at all.
"Half my interviews weren't getting detailed, structured feedback from panels," shares Anurag Tiwari, TA Head, Deutsche Telekom.
By the time a hiring decision needed to be made, the data was thin. The team was making calls on gut feel and half-finished notes. For a company hiring engineers at scale, that's an expensive way to operate.
Solution
Deutsche Telekom deployed JobTwine's AI Human Interviewer Copilot across their technical hiring rounds.
The first thing that changed was preparation — or rather, the elimination of it as a dependency. Each interview round got its own AI-powered playbook, built for that specific role and that specific stage. Panellists no longer needed to arrive with a plan in their heads. The plan was already there.
During live interviews, JayT — JobTwine's real-time AI Human Interviewer Copilot — guided panellists through the conversation as it happened. Suggested follow-ups. Tracked what had been covered. Flagged gaps. The interviewer could stay fully present — actually listening, actually probing — without the cognitive overhead of managing structure and evaluation simultaneously.
When the conversation ended, structured feedback went straight to the ATS. No chasing. No delays. No interpretation required.
JobTwine Products Implemented
Smart Playbooks
The first thing that changed was preparation — or rather, the elimination of it as a dependency. Each interview round got its own AI-powered playbook, built for that specific role and that specific stage. Panellists no longer needed to arrive with a plan in their heads. The plan was already there.
AI Avatar Recruiter
JayT conducted structured async screens before any live interview took place — giving panellists a complete, scored picture of each candidate ahead of the conversation. No cold starts. No relying on a CV alone.
AI Human Interviewer Copilot
During live interviews, JayT guided panellists through the conversation as it happened. Suggested follow-ups. Tracked what had been covered. Flagged gaps. The interviewer could stay fully present — actually listening, actually probing — without the cognitive overhead of managing structure and evaluation simultaneously. When the conversation ended, structured feedback went straight to the ATS. No chasing. No delays. No interpretation required.