Telemark Marketing: a CATI-room workload running on one configuration screen
Telemark runs a classic CATI workload on itellicoAI. What used to be a room of dozens of trained human callers now runs through voice agents — configured per study and reviewed in Quality Studio.
57%
completion rate
~2m18s
avg interview
DACH
accent coverage
24/7
sample window
Challenge
Company
CATI is one of the most labour-intensive workloads in the entire research industry. Hitting weekly quotas across multiple concurrent studies — public opinion polling, energy-market satisfaction, regional consumer studies, sports & leisure panels, transport-operator NPS, brand-recall — means staffing, training, and managing dozens of human callers. Quality varies caller-to-caller. Cost scales linearly with volume. Multi-region accents require local hires.
Market Research
Solution
Used Use Cases
itellicoAI runs each study as a distinct voice-agent configuration. The agent calls a respondent in their home region, opens with a regionally-natural greeting, explains the topic, conducts a 2–5 question structured interview with branching follow-ups, and writes a per-call goal-analysis verdict — 'interview completed,' 'interview started,' 'respondent declined,' etc. The same configuration model handles eight or more topic areas concurrently. Multiple sub-team accounts (NH, Shared, Master, Marketing variants) operate as a single, coordinated production environment.
Execution
Execution model
This case follows a practical rollout pattern that combines clear KPIs, integrated use cases, and weekly optimization.
Operational baseline
The team defined a clear KPI, mapped priority call intents, and agreed on transfer rules before launch.
Use-case deployment
Voice AI use cases were connected to the existing process so conversations could trigger real next actions.
Continuous optimization
Performance was improved with iterative prompt updates and regular review of call outcomes.
Conclusion
Telemark shows that structured market and opinion research works through Voice AI: studies are configured instead of coded, topics run in parallel, and regional accents are handled by the same model. The before/after is clean — a CATI room replaced by a configuration screen.
Want similar results?
We can map your process and launch a voice AI use case tailored to your team.
Questions
Case Study
FAQ.
What was the primary outcome?
57% completion rate.
Which use cases were used?
This deployment used Follow-up Calls, Survey Collection, Surveys & Feedback.
Can similar teams replicate this approach?
Yes. The same rollout pattern can be adapted to your call flows, goals, and escalation model.
How quickly can this be implemented?
Most teams launch an initial version in days and then optimize with live call data.
What is the best first step to replicate this?
Start with one high-impact call flow, define transfer guardrails, and measure one clear KPI from day one.