AI Economics Get Awkward in the Contact Center
Summary
Goldman Sachs research, cited on July 20, 2026, reveals a nuanced economic picture for AI agents, particularly in contact centers. While AI coding agents demonstrate significant cost savings at \$13.39 per day compared to \$300 for human developers, AI contact center agents are estimated at \$92.90 daily, slightly exceeding the \$90 cost of a human worker. This challenges the assumption that AI universally reduces operational expenses. The analysis stresses the importance of measuring the "whole outcome," considering factors like repeat contacts, customer satisfaction, and escalations, rather than just per-interaction cost. Consequently, the article suggests that AI's most effective near-term application in customer service may be augmentation, assisting human agents with complex tasks, which is projected to increase customer service software spending over the next two to three years, according to Opus Research's "The CCaaS Rebalance: CX Software Spend in the Agentic Era."
Key takeaway
For Directors of AI/ML evaluating customer service automation, do not assume AI agents are inherently cheaper than human counterparts. Your strategy should prioritize AI augmentation for complex interactions, helping experienced employees navigate systems and improve productivity, rather than solely focusing on replacing frontline staff. This approach, while not yielding dramatic workforce reductions, will likely deliver better service and a more economically sound automation strategy.
Key insights
AI agent cost savings are not universal; contact center AI can be more expensive than human agents, necessitating "whole outcome" measurement.
Principles
- AI cost savings vary by application.
- Measure AI value by "whole outcome."
- Augmentation can exceed full replacement.
In practice
- Assist agents in finding answers.
- Summarize customer conversations.
- Automate after-call work.
Topics
- AI Economics
- Contact Center AI
- Agent Augmentation
- Customer Service Automation
- Goldman Sachs
- Opus Research
Best for: CTO, Executive, Investor, Director of AI/ML, Consultant, VP of Engineering/Data
Related on AIssential
See Counsel's argued verdicts on the open AI decisions leaders are weighing →
Editorial summary, takeaway, and curation by AIssential. Original article published by Opus Research.