Agentic AI: A Comprehensive Survey of Technologies, Applications, and Societal Implications
Open accessThe DOI matches its Crossref record; 91 of the 91 references it cites were found in a registry or an index.
Abstract
Agentic AI brings a new level of advancement in artificial intelligence (AI), as it is capable of goal-directed behaviour, dynamic adaptation, and self-improvement. It influences various significant fields, such as robotics, healthcare, autonomous vehicles, and labor automation. This paper explores the defining features of agentic AI, highlights its differences from traditional AI, and discusses how autonomy, memory, goal-directed behavior, and adaptive reasoning contribute to increasingly general capabilities. Rather than proposing a new architecture, we offer a conceptual analysis of the trajectory from current agentic frameworks to future agentic AI systems. Examining applications in significant fields, including precision medicine, industrial robotics, and self-driving technologies, and discussing societal impacts, particularly concerning workforce disruption, augmentation, and ethical issues arising from agentic AI systems. The paper identifies key research directions for ensuring beneficial and controllable agentic AI development. This work serves as a primary source for exploring agentic AI by compiling existing research and identifying open questions.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“Despite architectural variation, all systems shared core agentic properties, autonomy, goal-directed execution, task decomposition, and tool interaction, supporting their classification as explainable agentic AI systems within clinical and biomedical contexts [2,5,28].”
From Explainable Agentic Artificial Intelligence in Healthcare: A Scoping Review (2026) · cites this paper as [28]
In the citing paper
“Although additional agentic characteristics were identified in individual studies, such as adaptability and memory [5,28], the heterogeneity of reporting and implementation made exhaustive enumeration impractical; therefore, we synthesized the most consistently reported and conceptually central features to provide a structured, comparable overview.”
From Explainable Agentic Artificial Intelligence in Healthcare: A Scoping Review (2026) · cites this paper as [28]
Citations
Cites
Edge computing: Vision and challenges
Shi, W., Cao, J., Zhang, Q., Li, Y., & Xu, L. (2016) · IEEE Internet of Things Journal · doi:10.1109/jiot.2016.2579198 · Cited by 8,733 · cited 1×
“Many researchers have tried to address these issues by introducing Hierarchical RL (decomposing complex tasks into small sub-tasks) - see [114], [115], Edge AI (computations are distributed to local nodes) - see [116], Federated learning(collaborative without raw data sharing) - see [117].”
Advances and Open Problems in Federated Learning
Kairouz, P., & McMahan, H. B. (2021) · Foundations and Trends in Machine Learning · doi:10.1561/2200000083 · Cited by 7,772
SOAR: An Architecture for General Intelligence
Laird, J. E., Newell, A., & Rosenbloom, P. S. (1987) · Artificial Intelligence · doi:10.1016/0004-3702(87)90050-6 · Cited by 2,535 · cited 1×
“SOAR is a rule-based architecture designed for general intelligence, integrating learning, problem-solving, and decision-making in a unified system; see [74], [75] and [76].”
Bellamy, R. K. E., Dey, K., Hind, M., Hoffman, S. C., Houde, S., Kannan, K., Lohia, P., Martino, J., Mehta, S., Mojsilovic, A., et al. (2018) · arXiv · doi:10.48550/arxiv.1810.01943 · Cited by 1,581
Computing Machinery and Intelligence
Turing, A. M. (2009) · Parsing the Turing Test · doi:10.1007/978-1-4020-6710-5_3 · Cited by 1,367
The Soar Cognitive Architecture
Laird, J. E. (2012) · doi:10.7551/mitpress/7688.001.0001 · Cited by 1,108
Imitation Learning: A Survey of Learning Methods
Hussein, A., Gaber, M. M., Elyan, E., & Jayne, C. (2018) · ACM Computing Surveys · doi:10.1145/3054912 · Cited by 1,060
Sim-to-Real Transfer in Deep Reinforcement Learning for Robotics: a Survey
Zhao, W., Queralta, J. P., & Westerlund, T. (2020) · 2020 IEEE Symposium Series on Computational Intelligence (SSCI) · doi:10.1109/ssci47803.2020.9308468 · Cited by 941 · cited 1×
“See [61], [62] Quantum computing: Quantum machine learning can help to learn complex patterns in data, quantum multi-agent reinforcement learning can help in scalability, and parallel and fast computing can be useful in agentic AI systems.”
AI Fairness 360: An extensible toolkit for detecting and mitigating algorithmic bias
Bellamy, R. K. E., Dey, K., Hind, M., Hoffman, S. C., Houde, S., Kannan, K., Lohia, P., Martino, J., Mehta, S., Mojsilović, A., et al. (2019) · IBM Journal of Research and Development · doi:10.1147/jrd.2019.2942287 · Cited by 884
Ahmed, I., Jeon, G., & Piccialli, F. (2022) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2022.3146552 · Cited by 853 · cited 1×
“To mitigate these challenges, researchers are integrating explainable AI (transparent and understandable AI decisions) - see [118], [119], interpretable cognitive architectures (human-like, transparent reasoning frameworks) -see [120], Neurosymbolic AI (which combines neural and symbolic reasoning)…”
Research works printed but not linked: 42
- Durante, 2024, Agent AI: Surveying the horizons of multimodal interaction, arXiv:2401.03568 · arXiv
- Gridach, 2025, Agentic AI for scientific discovery: A survey of progress, challenges, and future directions, arXiv:2503.08979 · arXiv
- Schneider, 2025, Generative to agentic AI: Survey, conceptualization, and challenges, arXiv:2504.18875 · arXiv
- Amine Ferrag, 2025, From LLM reasoning to autonomous AI agents: A comprehensive review, arXiv:2504.19678 · arXiv
- Chaudhari, 2025, Autonomous AI agents for real-time financial transaction monitoring and anomaly resolution using multi-agent reinforcement learning and explainable causal inference, Int. J. Advance Res., Ideas Innov. Technol. (IJARIIT), 11, 142
- Okpala, 2025, Agentic AI systems applied to tasks in financial services: Modeling and model risk management crews, arXiv:2502.05439 · arXiv
- Frenette, 2020, Automation and the sexes: Is job transformation more likely among women?, Anal. Stud. Branch Res. Paper Ser., 1, 1
- Kessavane, 2025, A study of reducing HR redundancy processes with agentic AI, IJAIDR-Journal Adv. Develop. Res., 16, 1
- Paul, 2024, AgentPeerTalk: Empowering students through agentic-AI-driven discernment of bullying and joking in peer interactions in schools, arXiv:2408.01459 · arXiv
- Bosch, 2024, AI chatbots as open educational resources: Enhancing student agency and self-directed learning, Italian J. Educ. Technol., 32, 53
- Mastrogiacomi, 2024, Facilitating a paradigm shift for teaching and learning with AIs, Italian J. Educ. Technol., 32, 69
- Tripathi, 2024, Revolutionizing customer service: How AI is transforming the customer experience, Amer. J. Comput. Archit., 11, 15
- Inavolu, 2024, Exploring AI-driven customer service: Evolution, architectures, opportunities, challenges and future directions, Int. J. Eng. Adv. Technol., 13, 156
- Patil, 2024, Artificial intelligence-driven supply chain optimization: Enhancing demand forecasting and cost reduction
- Wang, 2024, AI-driven day-to-day route choice, arXiv:2412.03338 · arXiv
- Katnapally, 2021, Automating cyber threat response using agentic AI and reinforcement learning techniques, J. Electr. Syst., 17, 138
- Ho, 2016, Generative adversarial imitation learning, Adv. Neural Inf. Process. Syst., 29, 4565
- Subramanian, 2024, A neuro-symbolic approach to multi-agent RL for interpretability and probabilistic decision making, arXiv:2402.13440 · arXiv
- Li, 2024, LogiCity: Advancing neuro-symbolic AI with abstract urban simulation, Adv. Neural Inf. Process. Syst., 37, 69840
- Lowe, 2017, Multi-agent actor-critic for mixed cooperative-competitive environments, Adv. Neural Inf. Process. Syst., 30, 6379
- Wang, 2021, AdderNet and its minimalist hardware design for energy-efficient artificial intelligence, arXiv:2101.10015 · arXiv
- Knott, 2021, Evaluating the robustness of collaborative agents, arXiv:2101.05507 · arXiv
- Florence, Implicit behavioral cloning, 158
- Rao, BDI agents: From theory to practice, 312
- Lanham, 2018, Learn Unity ML-Agents-Fundamentals of Unity Machine Learning: Incorporate New Powerful ML Algorithms Such as Deep Reinforcement Learning for Games
- Brockman, 2016, OpenAI gym, arXiv:1606.01540 · arXiv
- Palanisamy, 2018, Hands-On Intelligent Agents With OpenAI Gym: Your Guide to Developing AI Agents Using Deep Reinforcement Learning
- Kamalov, 2025, Evolution of AI in education: Agentic workflows, arXiv:2504.20082 · arXiv
- Jisna, 2025, Agentic AI: A comprehensive overview, Inst. Elect., 4, 1
- Scepanski, 2024, AI systems and their scalability—A systematic literature review, ACIS Proc., 2024, 1
- Singh, 2019, Employment transformation through artificial intelligence in India, Int. J. Appl. Eng. Res., 14, 65
- Frenette, 2020, Automation and job transformation in canada: Who’s at risk?, Anal. Stud. Branch Res. Paper Ser., 448, 1
- Ilyushina, 2022, Decentralised autonomous organisations: A new research agenda for labour economics, J. Brit. Blockchain Assoc., 5, 50
- Wang, 2024, Chain-of-Thought reasoning without prompting, arXiv:2402.10200 · arXiv
- Zhang, 2022, Automatic chain of thought prompting in large language models, arXiv:2210.03493 · arXiv
- Kolluri, 2024, An extensive investigation into guardians of the digital realm: AI-driven antivirus and cyber threat intelligence, Int. J. Adv. Res. Interdiscipl. Scientific Endeavours, 1, 71
- Kommera, 2024, Artificial intelligence in data integration: Addressing scalability, security, and real-time processing challenges, Int. J. Eng. Technol. Res. (IJETR), 9, 130
- 10.1561/9781680837896 · DOI · Paper page
- Evans, 2020, Kant’s cognitive architecture
- Ng, Algorithms for inverse reinforcement learning, 663
- Hadfield-Menell, 2016, Cooperative inverse reinforcement learning, Adv. Neural Inf. Process. Syst., 29, 3909
- Sutton, 1999, Reinforcement learning, J. Cognit. Neurosci., 11, 126
Cited by
Zhang, R., Liu, G., Liu, Y., Zhao, C., Wang, J., Xu, Y., Niyato, D. T., Kang, J., Li, Y., Mao, S., et al. (2026) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2026.3651702 · Cited by 148
“Frameworks including Codex [131], ReAct [70], and Toolformer [132] exemplify this transition through structured tool use and reasoning chains.”
Alva, L. R., & Pandey, B. (2026) · Artificial Intelligence Review · doi:10.1007/s10462-025-11458-6 · Cited by 32
“In applications such as autonomous driving and financial forecasting, more situation-aware and flexible agents were dis - covered when using utility-based agents (Ogbu 2023; Pati 2025).”
Agentic AI for smart and sustainable precision agriculture
Srinivasu, P. N., Pavate, A. A., JayaLakshmi, G., Shafi, J., Choi, J. I., & Ijaz, M. F. (2026) · Frontiers in Plant Science · doi:10.3389/fpls.2025.1706428 · Cited by 22
“Agents begin by collecting real-time data through sensors, imagers, and other input devices (Pati, 2025).”
Haidemariam, T. (2026) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2025.1728738 · Cited by 12
“The integration of agentic principles into artificial systems has profound socio-technical consequences (Pati, 2025).”
Pan, G., Tao, Y., Yang, Z., Li, J., Peng, Y., Bao, X., Zhang, Y., Hua, Z., Wang, S., Zhang, R., et al. (2026) · IEEE Transactions on Cognitive Communications and Networking · doi:10.1109/tccn.2026.3676045 · Cited by 10
“However, as we evolve towards the sixth-generation (6G) era, the network paradigm is shifting from merely connecting devices to connecting intelligence, giving rise to the internet of agents (IoA) [1], [2].”
Metadata
- Authors
- Pati, A. K.
- Affiliations
- Siksha O Anusandhan University (IN)
- Corresponding author
- Pati, Ashis Kumar
- ORCID
- Pati, Ashis Kumar
- Venue
- IEEE Access
- Published
- 2025
- Volume
- 13
- Pages
- 151824-151837
- ISSN
- 2169-3536 (ISSN-L)
- OpenAlex
- W4411995908
- Cited by
- 100
- References
- 87
- Citations by year
- 2026: 90 · 2025: 10
- FWCI
- 39.65
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Locations
- IEEE Access (publishedVersion, cc-by)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Artificial Intelligence Applications
- Physical Sciences › Computer Science › Artificial Intelligence › Psychiatry, Mental Health, Neuroscience
- Physical Sciences › Computer Science › Artificial Intelligence › AI-based Problem Solving and Planning
- Keywords
- agentic AI, goal-directed behavior, human development, autonomy, memory, adaptive reasoning, industrial robotics, precision medicine, autonomous vehicles, labor automation, workforce disruption, AI augmentation, AI ethics
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Agentic AI: A Comprehensive Survey of Technologies, Applications, and Societal Implications. https://app.cochat.ai/papers/10.1109/access.2025.3585609
- The paper
- Pati, A. K. (2025). Agentic AI: A Comprehensive Survey of Technologies, Applications, and Societal Implications. IEEE Access. https://doi.org/10.1109/access.2025.3585609
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.