E-ISSN:2583-2468

Research Article

Artificial Intelligence

Applied Science and Engineering Journal for Advanced Research

2025 Volume 4 Number 2 March
Publisherwww.singhpublication.com

Intent-Based Networking with AI: Towards Fully Autonomous Network Operations

Bairy V1*, Jorepalli S2
DOI:10.5281/zenodo.15347801

1* Vivek Bairy, Independent Researcher, San Francisco, United States of America.

2 Sunil Jorepalli, Independent Researcher, San Francisco, United States of America.

This research explored the integration of Artificial Intelligence (AI) into Intent-Based Networking (IBN) systems with the goal of enabling fully autonomous network operations. Utilizing a qualitative research design, the study involved expert interviews and system behavior observations within simulated environments. The AI-driven IBN prototype was developed with components for natural language intent recognition, policy translation, and autonomous fault management. The findings indicated high accuracy in interpreting user intents (93.7%), efficient policy deployment, and significant reductions in both configuration and recovery times compared to traditional Software-Defined Networking (SDN) systems. Experts validated the system's operational advantages while also noting challenges in handling ambiguous inputs and adapting to diverse network configurations. Overall, the research highlighted the feasibility and benefits of AI-enhanced IBN while recommending further real-world testing and security considerations to achieve truly autonomous network infrastructures.

Keywords: intent-based networking, artificial intelligence, network automation, natural language processing, policy translation, fault management, software-defined networking

Corresponding Author How to Cite this Article To Browse
Vivek Bairy, Independent Researcher, San Francisco, United States of America.
Email: 		Bairy V, Jorepalli S, Intent-Based Networking with AI: Towards Fully Autonomous Network Operations. Appl Sci Eng J Adv Res. 2025;4(2):39-44.
Available From
https://asejar.singhpublication.com/index.php/ojs/article/view/141

Manuscript Received Review Round 1 Review Round 2 Review Round 3 Accepted
2025-02-08 2025-02-28 2025-03-25
Conflict of Interest Funding Ethical Approval Plagiarism X-checker Note
None Nil Yes 2.86

© 2025 by Bairy V, Jorepalli S and Published by Singh Publication. This is an Open Access article licensed under a Creative Commons Attribution 4.0 International License https://creativecommons.org/licenses/by/4.0/ unported [CC BY 4.0].

Appl Sci Eng J Adv Res 2025;4(2)39
Download PDFBack To Article1. Introduction2. Literature
Review3. Research
Methodology4. Result and
Discussion5. ConclusionReferences
Bairy V., et al. Intent-Based Networking with AI

1. Introduction

The growing intricacy and necessity for high-performance networks have resulted in traditional methods of network management becoming less adequate. In opposition to this trend, Intent-Based Networking (IBN) presents itself as an exciting new paradigm that simplifies and automates network function through the ability to allow network managers to declare high-level business or operational intent, as opposed to detailed network settings. The abstraction to high-level intent offers much potential for simplifying network functioning, decreasing the need for human intervention, and increasing system flexibility overall.

With networks evolving toward next-generation technologies like 5G, 6G, and beyond, becoming increasingly more autonomous in terms of self-management capabilities and self-optimization capabilities is more critical than ever. The extension of Artificial Intelligence (AI) to IBN is expected to resolve these issues by further improving the network's capability to interpret, adapt, and enforce user intentions autonomously in real time. However, there are some challenges associated with AI implementation. By taking advantage of sophisticated AI methods like Natural Language Processing (NLP), machine learning, and reinforcement learning, AI-based IBN systems can potentially identify user goals independently, map them into network configurations, and dynamically adapt policies according to evolving conditions and user requirements.

The concept of completely autonomous network operations is the next milestone in the progression of IBN. It goes beyond the mere automation of chores to a mechanism in which the network can make intelligent decisions, predict faults, and recover without any help from humans. This transformation would help develop more efficient, scalable, and resilient networks that can deal with the complexity of applications today such as IoT, smart cities, enterprise cloud services, and AI-based applications.

Despite such encouraging possibilities, several issues do lie in getting to that kind of autonomy. They are assuring intent identification accuracy, maintaining strong fault detectability and recoverability, and assuring AI model adaptability in varying and dynamic network setups.

Further, effective implementation depends on surmounting challenges involved with data integrity, scalability, security, and interoperability.

This paper explores the role of AI in transforming IBN towards fully autonomous network operations, reviewing current methodologies, challenges, and advancements. Through both qualitative analysis and experimental results, it provides insights into the practical feasibility of deploying AI-driven IBN systems that can operate with minimal human oversight, thereby shaping the future of network management.

2. Literature Review

Wei, Peng, and Liu (2020) offered initial insights into the position of IBN within 6G networks. Their work highlighted that the complexity, heterogeneity, and user-centric nature of 6G service requirements made it a prime setting for the adoption of IBN. They presented various technical challenges, such as real-time intent recognition, context awareness, and adaptive orchestration, that had to be overcome to enable IBN to be realized to its full potential in future networks.

Leivadeas and Falkner (2022) surveyed IBN in breadth and depth, classifying current models, frameworks, and technologies. Their research came up with four key functions of IBN: intent capture, translation, validation, and assurance. The survey also pointed to the absence of standardization and the use of sophisticated AI techniques in enhancing system reliability and interpretability, especially in extensive deployments.

Andrade-Hoz, Wang, and Alcaraz-Calero (2024) added to the body of research by creating an infrastructure-wide dataset meant for training AI models in intent-based autonomous 5G networks. Their dataset filled the essential requirement for high-quality, labeled data to aid in the training and testing of AI components in real-world-like environments. The dataset further contained information across multiple network layers, allowing for multi-dimensional AI analysis to enhance intent inference and system behavior prediction.

Njah, Leivadeas, and Falkner (2024) suggested a complete AI-based IBN architecture that incorporated natural language processing (NLP), machine learning, and closed-loop automation.


Appl Sci Eng J Adv Res 2025;4(2)40
Bairy V., et al. Intent-Based Networking with AI

They illustrated in their architecture the potential of employing AI to sense, interpret, and enforce user intentions through dynamic and distributed network infrastructures. Nevertheless, they admitted that the system's performance was highly subject to the quality of training data as well as the resilience of the AI models against uncertain network scenarios.

Velasco et al. (2021) investigated the idea of end-to-end IBN, considering the orchestration of network services at multiple layers and domains. Their work emphasized that a single control plane that is capable of understanding the intents of users as a whole and applying policies consistently would be crucial. They also discussed interoperability, scalability, and intent conflict resolution as issues, which created substantial practical barriers to adopting IBN.

3. Research Methodology

3.1. Research Design

This research utilized a qualitative design to investigate the incorporation of Artificial Intelligence into Intent-Based Networking (IBN) systems with the aim of comprehending the possibilities and challenges of having completely autonomous network operations. The research was exploratory in nature and focused on collecting expert opinions and observational data using simulation environments. Through the focus on human opinions and behavioral results, the research aimed to interpret how AI-based technologies could transform conventional networking practices.

3.2. Data Collection Methods

Data was obtained mainly using participant observation and semi-structured interviews. Interviews were held with industry experts in a purposive sample, comprising network engineers, AI developers, and IT operations managers. The discussions were aimed at obtaining expert insights into existing network automation trends, the use of AI to enhance efficiency, and the perceived risk of implementing AI-based systems within production environments.

In addition, observational evidence was collected when developing and testing the AI-powered IBN prototype in a test network environment. Researchers recorded how the system reacted to different user intentions, coped with faults,

and evolved in response to topology changes. Observations were made of the system's performance as well as the responses of network operators interacting with the prototype.

3.3. Instrument Design and Procedure

Interview questions were open-ended and focused on key themes, including:

  • Experiences with conventional vs. AI-based network systems
  • Expectations and anxieties about AI automation
  • Observations regarding usability, trustworthiness, and control in autonomous systems
  • Recommendations for enhancing AI-intent recognition and fault handling

Each interview took around 45–60 minutes and was taped with permission. Answers were subsequently transcribed and examined for common themes and findings.

For observational data, there were comprehensive logs kept during test sessions. The logs contained network event descriptions, system responses, and subjective judgments by the observers of the system's performance and perceived autonomy.

3.4. Data Analysis Techniques

Qualitative data were processed through thematic analysis. Interview transcripts and observation notes were coded to determine recurring patterns, contradictions, and emerging themes. The important themes were: trust in AI-generated configurations, clarity and ambiguity in natural language commands, and the human perception of automation reliability.

Triangulation was employed to cross-validate information from various sources so that there would be a sound interpretation of the qualitative data. For example, interview feedback was compared with simulation session field notes to determine similar insights or contradictions in terms of system performance and usability.

4. Result and Discussion

The outcomes of the study were derived from robust testing of the AI-powered Intent-Based Networking (IBN) prototype in controlled settings and through qualitative data analysis of expert interviews.


Appl Sci Eng J Adv Res 2025;4(2)41
Bairy V., et al. Intent-Based Networking with AI

This section discusses findings on system precision, responsiveness, and fault handling and comparative assessments against conventional Software-Defined Networking (SDN) systems. The discussion reads these results within the context of realizing completely autonomous network operation, analyzing both strengths and present shortfalls of the AI-integrated solution.

4.1. Accuracy of Intent Recognition

The outcome shows that among 1,000 natural language test commands given, the system accurately resolved 937 commands, which amounts to an accuracy of 93.7%. This high accuracy proves the efficiency of the Natural Language Processing (NLP) model to comprehend user intentions and convert them into executable network configurations.

Table 1: Intent Recognition Accuracy

Total Test CommandsCorrectly InterpretedIncorrectly InterpretedAccuracy (%)
1,0009376393.7%

The 63 misinterpreted commands indicate cases where the system struggled, mostly from imprecise or ambiguous wording in the input. This indicates that though the system is good in most cases, there is potential for improvement in interpreting less explicit or context-based commands, which could be resolved by further fine-tuning the training data and model functionality.

4.2. Efficiency of Policy Translation

The findings show that among 950 tested policies, the system was able to translate 890 of them successfully, with a translation accuracy rate of 93.7%. This shows that the policy translation module, which is based on supervised learning, is very efficient in translating user intents into actionable network configurations.

Table 2: Policy Translation Accuracy

Total Policies TestedCorrect TranslationsIncorrect TranslationsTranslation Accuracy (%)
9508906093.7%

The 60 failed translations were mostly because of edge-case situations or new devices with non-standard configurations, which were challenging for the system. Even with such occasional errors, the general performance indicates that the translation engine can be trusted under most working scenarios,

and improvement is possible for more complex or changing network situations.

4.3. Response Time and Automation Efficiency

The findings indicate a considerable reduction in configuration deployment time using the AI-based Intent-Based Networking (AI-IBN) system compared to the conventional Software-Defined Networking (SDN) system. On average, AI-IBN deployed a new configuration in 3.2 seconds, whereas the SDN system took 7.6 seconds.

Table 3: Average Configuration Deployment Time

System TypeAverage Time (Seconds)Time Improvement (%)
SDN7.642.11%
AI-IBN3.257.9%

Figure 1: Average Configuration Deployment Time

This is a 57.9% decrease in deployment time, which reflects the effectiveness of the AI-IBN system. The quicker deployment is largely due to the AI model's capacity to create configurations in real-time, bypassing the need for manual translation processes and simplifying the process as a whole. This improvement in efficiency showcases the potential of AI-powered automation to greatly improve network operations and minimize latency.

4.4. Fault Detection and Recovery

The outcome proves decisively in favor of the AI-driven Intent-Based Networking (AI-IBN) system compared to conventional Software-Defined Networking (SDN) in the areas of fault detection and recovery.

Table 4: Fault Detection and Recovery Performance

MetricAI-IBNTraditional SDN
Fault Events Simulated5050
Automated Recoveries4715
Recovery Rate (%)94%30%
Average Recovery Time (sec)4.512.3

Figure 2: Fault Detection and Recovery Performance

Under simulation of 50 fault occurrences, AI-IBN automated recovery successfully in 47 cases with a recovery rate of 94%, which was considerably higher than the SDN system that managed to automate only 15 recoveries, for a recovery rate of 30%. Also, AI-IBN had a quicker average recovery time of 4.5 seconds, while SDN had a recovery time of 12.3 seconds.


Appl Sci Eng J Adv Res 2025;4(2)42
Bairy V., et al. Intent-Based Networking with AI

These results highlight the enhanced fault tolerance and fast self-healing nature of AI-IBN, which can significantly improve network reliability and decrease downtime by lessening manual intervention requirements.

4.5. Scalability Testing

Upon testing under different network topologies (10 to 500 nodes), the AI-IBN system performed steadily in intent interpretation and configuration deployment. The system recorded a mere 5–7% rise in processing time as network size increased tenfold, indicating remarkable scalability features.

4.6. Discussion

The findings evidently illustrated the capability of AI-boosted IBN to automate important network operations. The high accuracy achieved in both intent recognition and policy translation confirmed the viability of natural language-based network management. The drastic reduction in deployment and restoration times further underscored the advantages of real-time AI-based decision-making in mission-critical infrastructure.

In comparison to legacy SDN setups, the AI-IBN prototype demonstrated significant enhancements in responsiveness, fault tolerance, and autonomy. These results validated the hypothesis that integrating AI could make fully autonomous network functions possible. Improvement areas were also noted, including improved management of ambiguous commands and policy adaptation for infrequent or changing device settings.

Scalability outcomes were encouraging, indicating that the system can be stretched to the reach of service provider or enterprise-scale networks with little degradation in speed. However, the future works should be towards real-world deployments as well as hardening against security, especially in hostile environments where AI-driven decision-making can be manipulated.

5. Conclusion

In the incorporation of Artificial Intelligence into Intent-Based Networking showed great promise in making network operations completely autonomous systems. The AI-IBN prototype exhibited great accuracy in intent identification and policy conversion, high responsiveness in automating tasks, and strong fault management capabilities—all of which represented clear improvements from

conventional Software-Defined Networking. Qualitative feedback from the expert interviews supported the usability and efficacy of the system and revealed areas for further improvement, specifically in the management of ambiguous inputs and guaranteeing robustness against edge-case inputs. Collectively, these findings supported the conclusion that AI-powered IBN has great potential for smart, efficient, and scalable network management within dynamic, challenging environments.

References

1. Andrade-Hoz, J., Wang, Q., & Alcaraz-Calero, J. M. (2024). Infrastructure-wide and intent-based networking dataset for 5G-and-beyond AI-driven autonomous networks. Sensors, 24(3), 783.

2. Brodimas, D., Trantzas, K., Agko, B., Tziavas, G. C., Tranoris, C., Denazis, S., & Birbas, A. (2024, June). Towards intent-based network management for the 6g system adopting multimodal generative AI. in Joint European Conference on Networks and Communications & 6G Summit (EuCNC/6G Summit), pp. 848-853. IEEE.

3. Clemm, A. (2021). Intent-based network management.future networks, services and management: Underlay and overlay, edge, applications, slicing, cloud, space. AI/ML, and Quantum Computing, 383-398.

4. Leivadeas, A., & Falkner, M. (2022). A survey on intent-based networking. IEEE Communications Surveys & Tutorials, 25(1), 625-655.

5. Leivadeas, A., & Falkner, M. (2023, July). Autonomous network assurance in intent based networking: Vision and challenges. in 32nd International Conference on Computer Communications and Networks (ICCCN), pp. 1-10. IEEE.

6. Manias, D. M., Chouman, A., & Shami, A. (2024, May). Towards intent-based network management: Large language models for intent extraction in 5g core networks. in 20th International Conference on the Design of Reliable Communication Networks (DRCN), pp. 1-6. IEEE.

7. Mehmood, K., Kralevska, K., & Palma, D. (2023). Intent-driven autonomous network and service management in future cellular networks: A structured literature review. Computer Networks, 220, 109477.


Appl Sci Eng J Adv Res 2025;4(2)43
Bairy V., et al. Intent-Based Networking with AI

8. Mihaeljans, M., Skrastins, A., & Porins, J. (2024). Paramounts of Intent-based Networking: Overview. Elektronikair Elektrotechnika, 30(6), 53-59.

9. Njah, Y., Leivadeas, A., & Falkner, M. (2024). An AI-driven intent-based network architecture. IEEE Communications Magazine.

10. Pang, L., Yang, C., Chen, D., Song, Y., & Guizani, M. (2020). A survey on intent-driven networks. IEEE Access, 8, 22862-22873.

11. Sophocleous, M., Lessi, C., Xu, Z., Špaňhel, J., Qiu, R., Lendinez, A., ... & Belikaidis, I. (2022, June). AI-driven intent-based networking for 5G enhanced robot autonomy. in IFIP International Conference on Artificial Intelligence Applications and Innovations, pp. 61-70. Cham: Springer International Publishing.

12. Szilágyi, P. (2021). I2 BN: Intelligent intent based networks. Journal of ICT Standardization, 9(2), 159-200.

13. Velasco, L., Signorelli, M., De Dios, O. G., Papagianni, C., Bifulco, R., Olmos, J. J. V., ... & Ruiz, M. (2021). End-to-end intent-based networking. IEEE Communications Magazine, 59(10), 106-112.

14. Wang, J., Zhang, L., Yang, Y., Zhuang, Z., Qi, Q., Sun, H., ... & Liao, J. (2023). Network meets ChatGPT: Intent autonomous management, control and operation. Journal of Communications and Information Networks, 8(3), 239-255.

15. Wei, Y., Peng, M., & Liu, Y. (2020). Intent-based networks for 6G: Insights and challenges. Digital Communications and Networks, 6(3), 270-280.

Disclaimer / Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of Journals and/or the editor(s). Journals and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Appl Sci Eng J Adv Res 2025;4(2)44