The spring 2024 release from AIOps platform provider Selector aims to improve IT operations through generative artificial intelligence and advanced analytics. It also simplifies incident management, monitoring, troubleshooting, and planning.

Unlike traditional monitoring and observability tools, which focus on data collection and long-term storage, Selector analyzes operational telemetry to provide actionable insights. The platform predicts potential issues, such as network congestion or fiber optics overheating, enabling early intervention. According to Kevin Kamel, VP of product management, this starkly contrasts against other tools that retain data for years without providing immediate value.

Incorporating conversational AI

Founded in 2019 by former Juniper Networks executives, Selector initially targeted service providers but is now expanding into the enterprise sector. The startup recognizes the complex needs of enterprises to understand how network issues affect their infrastructure and applications. It takes a proactive approach to IT operations by continuously analyzing telemetry data and integrating conversational AI for more structured incident management.

“Integrating conversational AI allows us to field questions directly from the team where everybody can see the responses and work together to create a shared understanding of the incident and where things may have gone sideways,” Kamel said.

Selector specifically targets the challenges of tool sprawl within network operations centers (NOCs). Traditional setups in NOCs vary greatly among companies and often rely on ad hoc communication methods like phone calls, emails, or ticket systems. Using conversational AI, Selector facilitates a more structured approach to incident management, allowing team members to ask questions and receive responses in a shared space.

The company’s platform aggregates telemetry and metadata from various sources into its own data lakes, including metrics, logs, and events. It then applies AI and machine learning to analyze this comprehensive dataset. Kamel noted the platform isn’t just a set-and-forget system; it’s designed to adapt to the users’ specific needs.

The spring 2024 release addresses many use cases, such as observability, data center tool replacement, and digital twin creation. Here’s a closer look at the key highlights.

A leap forward with generative AI & analytics

At the heart of the new release is the integration of generative AI with Selector Copilot, a tool that converts incident summaries into actionable insights. Copilot provides IT teams with guided troubleshooting steps and creates summaries that are easier for humans to understand. This makes the platform more intuitive and accelerates the resolution of IT issues.

Previously, interactions with these tools relied on natural language processing (NLP), using Python libraries to translate user queries into a form the system could understand. Now, Selector has enhanced Copilot by integrating a large language model (LLM) and incorporating generative AI capabilities for summarization.

Selector has also improved root-cause analysis with a new feature that quickly identifies the underlying cause of an incident. It uses ML-based models to analyze the sequence and context of events. It also incorporates domain expertise, improving the platform’s ability to interpret ambiguous data and provide more precise insights into network behavior and potential issues.

Over the decades, multiple vendors have attempted to improve root-cause analysis with little success. One of the challenges is there is far too much data to process promptly. I view AI as the missing piece to make a root cause NOC tool with the degree of accuracy required by IT pros today.

Enhanced monitoring and observability

The update substantially boosts Selector’s monitoring and observability capabilities, enabling comprehensive data collection and analysis across network infrastructures and applications. The platform has evolved from collecting, storing, and analyzing telemetry data. It now includes service-level monitoring capabilities, enabling users to manage service-level objectives (SLOs) and service-level agreements (SLAs). The addition marks a significant shift, as traditionally, the network domain hasn’t focused much on SLOs and SLAs.

The introduction of service-level monitoring allows the platform to replace various legacy systems and offer additional value through AIOps and enhanced service-level oversight. According to Kamel, this approach has replaced numerous tools within the platform’s existing customer base, demonstrating its effectiveness and the added benefits it brings to network management and operations.

Digital twin creation for network & IT infrastructure

A new feature in this release allows users to design and use digital models of their network and IT infrastructure. This feature provides visualizations that help users explore traffic flows and routing changes, facilitating better decision-making.

The digital twin capability extends beyond simply reviewing past network configurations and events. It can simulate future scenarios by manipulating network elements to see how changes would affect traffic flow based on historical data. For instance, users can explore the consequences of removing a network link between two nodes, including the impact on traffic volume, latency, and jitter. The system tracks the redundancy of network paths to understand how traffic reroutes and whether the alternative paths can handle the increased load.

Given the cost of network equipment and the pace of change, few companies can build a full-scale lab to try things out before putting them into production. A digital twin of the network lets companies run what-if scenarios, experiment with network settings, and confirm updates before moving them into production.

Simplified access through GCP marketplace

The platform is expanding into cloud marketplaces, starting with the Google Cloud Platform (GCP) marketplace, which plans to launch in early Q2. This move is part of a strategic pivot toward the enterprise sector, which prefers operational expenditures (opex) over capital expenditures (capex). Marketplaces offer customers a hassle-free way to procure software, simplify vendor management, and list pricing options.

Unlike service providers, enterprise IT departments often rely exclusively on cloud services. The GCP marketplace caters to this preference, enabling customers like large hotel chains to easily discover and access AIOps solutions. Customers can choose a proof of concept or enter into multiyear agreements directly through the marketplace.

The takeaway

The Selector’s spring 2024 release advances IT operations management by combining generative AI, advanced analytics, and user-centric data. Selector addresses many of the current pain points for network engineers of service providers and large enterprises.

Networks are growing in complexity, and AI is the only path to simplicity. Generative AI brings natural language, which furthers the simplicity of operating a network.

“We’re in the right place at the right time with this. There’s been an overwhelming interest in the product,” said Kamel. “Assistive troubleshooting, automated incident remediation, and incident summarization are key features that operations leaders have been demanding for years. Folks are showing up and saying: ‘We need this!’”

Zeus Kerravala is the founder and principal analyst with ZK Research. Prior to ZK Research, Kerravala spent 10 years as an analyst at Yankee Group.