Podcast Summary: Using AI at Work

Episode 97: Using AI for Customer Support – Voice AI vs Humans in Customer Service Strategy with Nathan Strum

Date: March 30, 2026
Host: Chris Daigle
Guest: Nathan Strum, CEO & Co-Founder of Abby Connect

Overview

This episode delves into the evolving landscape of customer support, specifically the roles of Voice AI and human agents in business communication. Chris Daigle interviews Nathan Strum, who blends his two decades of experience in human-led customer service with Abby Connect’s expansion into AI-powered solutions. Together, they examine the present and future of AI in customer support, the dynamics of AI adoption, ethical considerations, and strategies for thoughtfully integrating AI without sacrificing the human touch.

Key Discussion Points and Insights

1. Why Human Agents Remain Essential

• Edge Case Handling:
  • Nathan: “Voice agents are very good at sticking to scripts, … but you can never get rid of edge cases, and unpredictability is best handled by humans.” [00:10]
• Customer Preferences:
  • Chris: “If it would be the same effort to work with an AI or work with a person, my preference would be to work with the person.” [04:56]
• Conversion Rates:
  • Humans are better at converting new clients or prospects compared to AI, which can miss nuances in sales or complex issues. [07:38]

2. Where Voice AI Excels

• Repetitive, Structured Tasks:
  • Voice AI shines in scenarios like appointment scheduling, routing calls, and answering within one ring — especially when a live receptionist is cost-prohibitive.
  • Nathan: “Within one ring, calls are answered. People feel like they’re heard.” [16:00]
• Scalability and Cost:
  • AI can handle routine inquiries 24/7 and free up human agents for more complex interactions.

3. Pitfalls, Risks, and Ethical Considerations

• Outbound Calling with AI:
  • Both Nathan and Chris caution strongly against using AI for cold outbound calls due to high risk, poor reception, and legal scrutiny.
    • Nathan: “It’s already a risky job to be doing an outbound call…You throw AI on top of that, you’re just lighting the fuel on fire.” [15:44]
• Transparency with Callers:
  • Disclosing the use of AI at the outset dramatically improves caller satisfaction and trust.
    • Nathan: “From an ethical standpoint, yes, you should in every scenario… announce it’s AI.” [17:31]
    • If callers feel deceived, trust erodes quickly.
• Legal Compliance:
  • No current US federal laws require disclosure, but legislation is under debate and best practices recommend transparency. [21:48]

4. Human vs. AI: Real-World Examples & Trends

• Case Studies (Klarna):
  • Chris shares how Klarna initially replaced customer service staff with bots and achieved equal customer satisfaction scores, but later reversed course and rehired humans.
    • The lesson: quantitative metrics alone don’t reflect deeper customer experience. [09:09–11:37]
• Abby Connect’s Experience:
  • Some Abby customers tried AI for cost reasons but switched back after negative caller feedback — supporting the need for human escalation paths. [11:04]

5. Implementation & Continuous Improvement

• AI Adoption at Abby Connect:
  • Timeline: First beta product in 2024, official launch in 2025 (leveraging tech like OpenAI, Eleven Labs).
    • Big advances in speech-to-text and accent detection are enabling broader use cases. [12:23–14:21]
• Change Management:
  • Abby introduced AI to staff through personal use cases, training modules, and transparent communication to ease anxiety around job loss. [29:56–31:50]
  • Employees are upskilled to adapt to an AI-augmented environment.

6. The Future: Hybrid Approaches & Market Disruption

• Hybrid Models:
  • Preferable strategy is to deploy AI for routine calls and reserve humans for complex, sensitive, or high-value interactions.
  • Nathan: “There are fundamentally conversations that just need to be between one human and another human.” [26:10]
• Offshore & Accent Modification Tech:
  • Offshoring is still prevalent, but accent modification AI is being adopted, though not yet widespread. [23:45–24:18]
• Impact on Jobs:
  • Despite dire predictions (e.g., Anthropic’s forecast that 50% of entry-level white collar jobs will be automated), Nathan is skeptical and instead sees evolving roles rather than simple replacement. [25:39–26:09]

Notable Quotes & Memorable Moments

• On Human Superiority in Edge Cases:
  • “Unpredictability is best handled by a human.” — Nathan [00:10, 04:21]
• On AI Disclosure:
  • “The results were just incredibly better when you announce it’s AI.” — Nathan [17:31]
  • “Imagine the trust out the window…especially a company you put your lives in their hands.” — Nathan (on undisclosed AI at airlines) [20:11]
• On Cautious Use of AI for Outbound Calls:
  • “Please don’t do outbound calls with AI… You’re just lighting the [fuel] on fire.” — Nathan [15:23, 15:44]
• On Upskilling Employees:
  • “We introduced learning management software modules for you to learn AI so that you can develop the skills necessary to get the new jobs in this new society.” — Nathan [30:04]
• On AI’s True Role:
  • “We don’t look at AI as a cost cutter. We look at it as rebalancing expenses.” — Nathan [27:29]

Timestamps for Important Segments

• 00:10 | Why humans beat AI at edge cases in customer service
• 06:52 | Scheduling as a clear win for Voice AI; sales as a human task
• 09:09 – 11:37 | Klarna case study — replacing/reinstating human customer service
• 12:23 | Abby’s timeline for AI adoption and key tech drivers
• 15:23 | Importance of not using AI for outbound voice sales/cold calling
• 17:31 | Ethics and best practices around disclosing AI to callers
• 21:48 | Legal/regulatory environment for AI in the US, current state
• 26:10 | Nathan’s core belief: some conversations should be exclusively human
• 27:29 | How Abby reframes AI as an expense rebalance, not a mass layoff tool
• 29:56 | Change management and staff transparency when introducing AI
• 33:30 | Industries and best-fit clients for Abby Connect
• 35:52 | Warning about purely tech-oriented vendors neglecting human service

Best Practices and Recommendations

• Never use AI for outbound cold calls due to legal, ethical, and customer-experience risks.
• Always provide a clear, easy escalation path to a human agent when deploying AI.
• Proactively disclose AI presence to maintain trust and set expectations.
• Deploy AI for simple, routine inbound tasks where speed and consistency matter (e.g., scheduling, basic information, after-hours response).
• Review customer feedback and monitor for any drop-off in satisfaction—be ready to revert or hybridize the model as necessary.
• Approach AI integration as a tool for augmenting and empowering staff, not just cutting costs.
• Invest in continuous team training and transparent communication to ease transition anxiety.

Conclusion

Nathan Strum and Chris Daigle provide a nuanced, experience-based blueprint for blending Voice AI with human customer service. The key: Embrace AI’s efficiency and scalability for routine interactions, but always retain and value the irreplaceable edge that humans bring to complex, unpredictable, or sensitive customer needs. Thoughtful integration—grounded in ethics, transparency, and flexibility—is essential for sustainable customer experience and team wellbeing.

Learn more about Abby Connect at abby.com and follow them on LinkedIn (“Abby Connect”).