Artificial intelligence (AI) has enabled consumer applications from speech and face recognition to self-learning IoT platforms, while automation allows for positive feedback loops within organizational and operational processes. While both these technologies will develop individually, it’s their synergies that will make the biggest impact. AI-assisted automation (AA) has the power to revolutionize the way we live and work today. Consequently, industries across the spectrum need to understand the opportunities, and undoubtedly the responsibilities, in light of the currently available AA landscape. As we enter 2018, let’s have a look at the following list of trends and predictions for the networking sector.

AI and automation synergies

  • Streaming telemetry: 2018 will witness advanced telemetry platforms coming into existence. The monitoring technologies will evolve beyond proprietary polling-based mechanisms to data streaming for end-to-end network probing. This will enable data-driven automation with underlying AI algorithms.
  • Predictive and prescriptive analytics: Moving away from threshold-triggered alarm systems, cognitive operational assurance architectures have already surfaced last year. This year the boundary between predictive (e.g., proactive alerts) and prescriptive (action) analytics will fade away, allowing for self-improving network operations.
  • Zero touch networking: The promise of automated network design, deployment and optimization at scale will become a reality, with the first multi-domain, multi-layer cognitive management prototypes and demonstrations being carried out this year.
  • Operator experience: While user-experience is routinely touted as the ultimate market differentiator, it exclusively refers to end-user applications and services, with little attention to underlying stakeholders. AA-driven data consumption and insights are expected to transform business and operational initiatives via self-service platforms, drag-and-drop service management, etc. enabling a unique network operator experience.
  • Network security: Until last year, most of the network security aspects were confined to isolated network layers. As more value is assigned to data and underlying infrastructure, 2018 will see multi-layer AA-driven security frameworks, capturing continuously evolving reference network behaviour.

Together with hot networking applications in 2018, it’s equally important to highlight underlying technology trends in the coming year.


  • Deep reinforcement learning: Deep reinforcement learning was the standout technology in 2017 with the success of AlphaGo Zero learning by playing against itself, countless start-up announcements, etc. As the war of deep learning frameworks settles down, 2018 will see more focus on generalized learning theory and related ensemble model developments.
  • Blockchain-enabled distributed analytics: Blockchain has been the key driver behind crypto currencies. Blockchain may fundamentally incorporate user data control via secure access by distributing data in blocks (as opposed to centralizing it at a single vendor) and enable distributed analytics across multiple stakeholders.
  • Digital twin: The concept of a digital twin refers to a digital representation of a product or service based on non-stop asset monitoring and analysis. Originating in the manufacturing sector, it is expected to trickle across industries encompassing both physical and virtual resources.
  • Automated ML: The typical machine learning (ML) workflow is a cumbersome one, including data preparation, feature engineering, model selection, training, validation and further optimizations. Automated ML will be one of the hot trends in 2018, aiming to automate this workflow using ML itself. Interesting indeed!
  • Self-explanatory models: Most ML models are considered to be black boxes. While there are accuracy and performance scores available, how an outcome was achieved is largely unknown. 2018 will see more effort in the direction of transparent and unbiased ML, establishing trust in the technology and leading to wider adoption.

The coming year will witness AI-assisted automation to surpass other networking trends. Equipment vendors and operators alike will embrace AA-first policies to ensure optimized network operations, drive-down costs and ensure better flexibility than current software suites can offer. AA will move from a nice-to-have to a must-have.