CCAIA Chicago Meeting 05/29

Help choose sub-topics for the Chicago meeting on Voice Agents

May 05, 2026

The Contact Center AI Association’s fifth meeting of the year is in

Chicago on May 29, 2026 from 2pm to 5pm (Central) at Ulta Beauty Guest Services. 

This meeting topic is:

Deploying Generative Voice AI

Association members have expressed the following sub-topics as most important. Please help us refine the agenda by letting us know what’s most important to you.

Please review the topics below and reply to share your suggestions:

  1. Which team should own generative voice AI? IT? CX? Product? Operations?

  2. Pre-launch QA & testing frameworks: What processes should be in place before a voice AI agent goes live? Simulation, staging environments, and test coverage. Defining “ready for production” in voice AI.

  3. Post-QA optimization & continuous improvement: What actually happens after quality reviews? Turning QA insights into model, prompt, and flow improvements. Closing the loop between evaluation → iteration.

  4. Human evals versus judges: When should humans evaluate vs LLM-based judges? What’s the right balance? And designing effective feedback loops and scoring systems.

  5. Bot-to-bot testing & synthetic users: Who is using automated agents to test other agents? Scaling QA with synthetic conversations. Pros/cons vs real-user testing.

  6. Observability - actions vs conversations: Why tracking actions taken matters more than just transcripts. Instrumentation for decision-level visibility. Debugging beyond “what was said.”

  7. Language detection & multilingual reliability: Preventing mid-call language switching. Role of LID (language identification) models. Latency vs accuracy tradeoffs in preprocessing. Prompting vs system-level enforcement for language control.

  8. Voice AI infrastructure pitfalls (beyond “just prompts”): Why system prompts alone aren’t enough. Importance of upstream components (LID, ASR accuracy, latency). Hidden failure points in the voice stack.

  9. AI for sentiment analysis, coaching & next-best-action: Using AI to detect human sentiment in real time. Impact of EU AI Act on these use cases.

  10. Regulation & compliance in voice AI: Risk levels for sentiment analysis and behavioral AI. Designing compliant AI-driven CX systems.

  11. Defining quality & success in voice AI: Task success vs experience vs business outcomes. What “good” looks like at different maturity stages. Aligning QA metrics to business KPIs.

  12. Evaluation metrics & measurement systems: Beyond WER: task completion, latency, containment. Measuring soft skills (empathy, tone, resolution quality). Conversation-level vs action-level scoring.

  13. Knowledge base & grounding strategy: Retrieval quality and hallucination prevention. Structuring knowledge for voice vs chat. Vocabulary/domain control.

  14. Bot behavior management & control systems: Prompting vs orchestration vs guardrails. Managing edge cases and escalation paths. Preventing unintended behaviors.

  15. System architecture & integration challenges: Connecting backend systems. Workflow orchestration and dependencies. Why integration is the main failure point.

  16. Adoption & change management: Bridging the gap between capability and real usage.

  17. AI drift & lifecycle management": Detecting regression over time. Versioning, testing, and rollout strategies. Maintaining quality post-deployment.

  18. ROI & business impact: Linking QA metrics to outcomes (CSAT, containment, cost). Cost vs performance tradeoffs. Proving value of voice AI investments.

Respond to this email to let us know if you like these sub-topics, highlight your top priorities, or recommend alternatives.

Express interest in attending in person or virtually through this form.

Meeting capacity is limited, so grab your seat.

Our meeting environment is established on a membership-priority basis.
We highly encourage you to join today!

Follow us on LinkedIn to stay up to date.