Daniella's Customer Brain

Surpless Dunn <> OneAI: Agent Updates

3 min read

Surpless Dunn <> OneAI: Agent Updates

Date: 2026-02-18 | Duration: 37.52000045776367 min | Account: surpless-dunn

External attendees: kknight@surplessdunn.com, pnelson@surplessdunn.com, wpinnello@surplessdunn.com

Summary

During the meeting, the team discussed the inefficiencies of the current AI agent scripts, which were deemed too complex. Kevin Knight highlighted the need for simpler, direct scripts that facilitate quick qualifications and customer transfers to live agents. Tests on a simplified ‘Need Help?’ script showed differing qualification rates, prompting Daniella Block to emphasize iterative testing. Patti Nelson and Bill Pinnello advocated for even more straightforward approaches in interactions, suggesting upfront identification of the AI and quick transfer offers. The importance of maintaining brand trust and improving call handling processes to ensure smooth transitions and address customer confusion was underscored by several participants. The team committed to a phased testing approach, with a focus on transparent adjustments based on real customer interactions.

Key Points

📝 AI Script Simplification: Current complex AI scripts decrease engagement; aim for simple, direct scripts for quick transfers to live agents. 📊 Testing Results: Simplified ‘Need Help?’ script had a 10% qualification rate; controlled rollouts help improve performance incrementally. 🤝 Direct Customer Engagement: Focus on identifying a customer’s interest upfront; simple messages improve clarity and connection rates. 📞 Call Handling Improvements: Address call repeats and confusion during customer holds; accurate responses are vital for maintaining trust. 🔄 Iterative Testing Approach: Phased testing with 20-40 call batches to monitor reactions and make adjustments based on real interactions. 🏆 Brand Trust Importance: Preserve customer relationships by ensuring AI calls do not frustrate; clear communication about AI usage is essential.

Overview

  • AI Script Simplification: Current complex AI scripts decrease engagement; aim for simple, direct scripts for quick transfers to live agents.
  • Testing Results: Simplified “Need Help?” script had a 10% qualification rate; controlled rollouts help improve performance incrementally.
  • Direct Customer Engagement: Focus on identifying a customer’s interest upfront; simple messages improve clarity and connection rates.
  • Call Handling Improvements: Address call repeats and confusion during customer holds; accurate responses are vital for maintaining trust.
  • Iterative Testing Approach: Phased testing with 20-40 call batches to monitor reactions and make adjustments based on real interactions.
  • Brand Trust Importance: Preserve customer relationships by ensuring AI calls do not frustrate; clear communication about AI usage is essential.

Action Items

Daniella Block Review specific call recordings identified by Kevin Knight to investigate repeated calls and agent response issues (16:37) Prepare and share test recording of the proposed simplified script for client approval before next rollout (35:13) Implement AI technical fixes currently underway based on analyst feedback and call data (10:18) Coordinate next controlled rollout test with ~20-40 calls using agreed simplified script (34:59)

Kevin Knight Provide Daniella Block with links to specific calls illustrating repeated calling issues for investigation (17:32)

All/Team Conduct two iterative rounds of testing with small call batches, analyze results to assess if AI outbound agent can meet success criteria or requires different approach (33:44)

Graph View