Author: Pawel Korzynski and 5 other multinational professorsJournal: European Management Journal (SSCI, IF=8.3)DOI: 10.1016/j.emj.2025.06.002English Title:Humanoid robotics and agentic AI: reframing management theories and future research directions
🧠 One-Sentence Preview
The “physical” humanoid robots and “cloud-based” AI agents are working together to dismantle the traditional management premise of “human-centered” —team roles, value co-creation, and lean six sigma all need to be rewritten!
🌪 Research Background: Management Theories Encountering the “Non-Human” Shockwave
-
Generative AI has swept through text and images, but **embodied AI (Physical AI)** has just emerged:– Humanoid Robots: Digit moving boxes at Amazon, Atlas jumping at Hyundai, Apollo in the Mercedes production line– AI Agents: OpenAI Operator, Byte’s UI-TARS, Synthesia digital humans… 24/7 autonomous decision-making
-
Traditional theories assume “humans are the protagonists”; when faced withmachines as social actors: – How to divide team roles?– Who co-creates value with the brand?– Do lean/six sigma still need human black belts?
🎯 Purpose of This Article
To systematically outline the core theoretical impacts of humanoid robots & AI agents on HRM / Marketing / Operations and propose three meta-rules: **”role redistribution – real-time response – bidirectional adaptation”**, providing a practical future research checklist.
🔍 Theoretical Breakdown & New Propositions
1️⃣ HRM: From “Person-Job Fit” to “Person-Machine Fit”
| Classic Theory | Traditional Assumption | New Context | Extended Propositions by Authors |
|---|---|---|---|
| Belbin Team Roles | 9 roles are played by humans | Robots can act as Trainer/Tester | Need to addmachine-oriented roles; humans need to “delegate” roles and then reclaim them |
| P-E Fit | Humans with environment (organization, position, group) | AddPerson-Machine Fit | Three sub-dimensions: communication style, complementary abilities, ethical culture |
| Social Learning Theory (SLT) | Humans observe “human” models | AI agents become24h conversational models | Introducebehavioral simulation + real-time feedback stages, formingreciprocal learning |
📌 One-Sentence Takeaway: In the future, HR will not only measure “person-job fit” but also “person-machine fit” — if robots speak too mechanically, employee P-M Fit will be low, and turnover will still be high!
2️⃣ Marketing and Consumer Behavior: Value Co-Creation Is No Longer “Human-Driven”
| Classic Theory | Traditional Logic | Impact of Robots/AI |
|---|---|---|
| Service-Dominant Logic (S-D Logic) | Value is co-created by “human-human” interactions | AI agents areautonomous value co-creators; canreal-time shapeconsumer value perception |
| Parasocial Relationships | Humans have one-way emotional ties to media | AI digital humans bringswift PSI (instant parasocial): real-time, bidirectional, dissipating once ended |
| Social Agency Theory | Humans hold agency | Humanoid robots’ appearance + algorithmic decision-making →consumer perception of agency transfer; if robots lead, customers will feel anxious |
📌 Experimental Evidence:
-
NYC police test Knightscope K5 robots for patrol —human officers must accompany, otherwise citizens’ sense of safety plummets (Shanks et al., 2024)
-
In restaurants,groups of customers prefer humanoid robots (social sense),while single customers prefer non-humanoid (to avoid feeling oppressed) (Huang & Liu, 2022)
3️⃣ Operations Management: Lean/Six Sigma/TOC Enhanced by AI
| Methodology | Traditional Pain Points | New Approaches with Robots/AI |
|---|---|---|
| Lean | Manual Kanban, post-hoc waste identification | AI agentsreal-time demand forecasting + automatic JIT; value stream mappingdynamic refresh |
| Six Sigma (DMAIC) | Humans take time to Define/Measure | AI can instantly capture social media complaints →automatically Define; IoT + AI 24h Measure; authors proposeDMALIC(+Learning) |
| Theory of Constraints (TOC) | Post-hoc identification of bottlenecks | Humanoid robotsreassign tasks on-site; shift loads to adjacent production lines 5 minutes early |
📌 Case Study:GPT industrial bot reducedoperational error rate by 38%,real-time adaptability increased by 52% (Kiangala & Wang, 2025)
🧭 Three Meta-Rules (Applicable Across Disciplines)
-
Redistribution of Roles and PowerHumans delegate part of decision-making/leadership → machines become Trainers/Leaders/value co-creators
-
Real-Time Response Becomes DefaultUpgrading from “weekly reports + meetings” to “second-level feedback + automatic adjustments”
-
Bidirectional Mutual AdaptationHumans change robots, and robots also change humans; organizations becomeco-evolutionary systems
🚀 Future Research Checklist (Ready for Proposal)
| Field | Specific Topics | Method Suggestions |
|---|---|---|
| HRM | When is machine leadership accepted? — Cross-cultural/task type experiments | 2×2 experiments (high/low power distance × routine/creative tasks) |
| Marketing | The higher the “appearance” of digital humans, the stronger the swift PSI? — Appearance vs. trust inverted U-shape | Eye-tracking + emotional facial analysis |
| Operations | How early should humanoid robots adjust loads in TOC for optimal performance? — Digital twin simulation | AnyLogic/Plant Simulation |
| Interdisciplinary Ethics | Robots collecting facial/voice data 24h → employees/customersWhere are the privacy boundaries? | Scenario-based questionnaires + GDPR compliance design |
⚠️ Ethical Alarm
-
**Lack of “Corporate Digital Responsibility (CDR)” can lead to:– Algorithmic bias → unfair recruitment/recommendations– Over-monitoring → decreased employee dignity
-
Deep voice cloning → customers being deceived by “fake employees”
-
The authors call for establishingthree layers of governance:① Cultural layer (shared norms + CDR KPI)② Structural layer (Chief Ethics Officer + cross-departmental committee)③ Process layer (data minimization + explainable AI + human fallback)
📖 One-Sentence Conclusion
Do not treat humanoid robots and AI agents as “advanced tools” — they arereshaping organizational structures, power dynamics, and value logic; management scholars and practitioners mustactively rewrite theories and design rules in advance, to gain a competitive edge in the era of “human-machine co-performance”.
END
Click to follow, send “Consumer behavior” to obtain the full PDF