Seminar & Forum
Academic SeminarIs AI an Invader That Threatens Humans? Complementary Effects between Agents and AI on the Productivity of Call center Agents
- Date
- 2019-03-28 ~ 2019-03-28
- Time
- 16:00 ~ 17:30
- Place
- Supex Building, Lecture Room 101
- Department
- School of Management Engineering
- Manager
- Seulbi Lee , seulbilee@kaist.ac.kr
- Major
- IT Management
We would like to invite you to participate in Management Engineering(ME) Seminar.
1. When: March 28th (Thursday), 16:00~17:20
2. Where: Supex Building, Lecture Room 101
3. Speaker: Prof. Seung Hyun Kim (Yonsei University)
4. Topic: Is AI an Invader That Threatens Humans? Complementary Effects between Agents and AI on the Productivity of Call center Agents
5. Research field: IT Management
* Lecture will be delivered in Korean.
* Seminar materials: Abstract
Abstract:
The promises of artificial intelligence (AI) co-exist with a skepticism that AI-based systems can live up to them. In this study, we adopted a perspective of learning curves and examined (1) the impact of introducing AI-based systems on the task performance of call center agents and (2) the conditions that moderate the impact of using AI-based systems. By analyzing the panel data of 360 agents, we first found a short-term loss of productivity, possibly because making direct inquiries to the AI-based systems during a call may drain an agent’s cognitive resources and be a distraction. In contrast, we found that a higher cumulative usage of direct inquiries to AI improved agents’ productivity. Interestingly, we observed a time reallocation effect as an additional benefit of prior usage of AI by which agents with this prior experience took less time on the phone to satisfy their customers and later spent slightly more time in postprocessing while the sum of the two is reduced. We further found that the impact of cumulative direct inquiries to AI depends on the characteristics of AI and of human agents. Specifically, this impact is amplified when AI-based systems are more accurate or have achieved a higher degree of learning. Moreover, this cumulative impact of direct inquiries to AI decreases with agents with higher IT savviness. Our results highlight the importance of learning by both human agents and AI-based systems, which helps overcome short-term losses in performance. Thus, for greater success, humans and AI should complement each other by reinforcing each other’s strengths.
1. When: March 28th (Thursday), 16:00~17:20
2. Where: Supex Building, Lecture Room 101
3. Speaker: Prof. Seung Hyun Kim (Yonsei University)
4. Topic: Is AI an Invader That Threatens Humans? Complementary Effects between Agents and AI on the Productivity of Call center Agents
5. Research field: IT Management
* Lecture will be delivered in Korean.
* Seminar materials: Abstract
Abstract:
The promises of artificial intelligence (AI) co-exist with a skepticism that AI-based systems can live up to them. In this study, we adopted a perspective of learning curves and examined (1) the impact of introducing AI-based systems on the task performance of call center agents and (2) the conditions that moderate the impact of using AI-based systems. By analyzing the panel data of 360 agents, we first found a short-term loss of productivity, possibly because making direct inquiries to the AI-based systems during a call may drain an agent’s cognitive resources and be a distraction. In contrast, we found that a higher cumulative usage of direct inquiries to AI improved agents’ productivity. Interestingly, we observed a time reallocation effect as an additional benefit of prior usage of AI by which agents with this prior experience took less time on the phone to satisfy their customers and later spent slightly more time in postprocessing while the sum of the two is reduced. We further found that the impact of cumulative direct inquiries to AI depends on the characteristics of AI and of human agents. Specifically, this impact is amplified when AI-based systems are more accurate or have achieved a higher degree of learning. Moreover, this cumulative impact of direct inquiries to AI decreases with agents with higher IT savviness. Our results highlight the importance of learning by both human agents and AI-based systems, which helps overcome short-term losses in performance. Thus, for greater success, humans and AI should complement each other by reinforcing each other’s strengths.
Contact : Lee, Jisun ( jisunlee@kaist.ac.kr )
