BigPanda Unveils Generative AI Tool Tailored for Enhanced ITOps Management

IT operations teams face a constant whirlwind of challenges, particularly when incidents occur that incapacitate essential systems. Time is always of the essence, and companies have long sought advantages to restore services quickly. To that end, many have developed playbooks for common issues and conduct postmortems to avoid future recurrences. However, not every problem has a straightforward solution, especially given the vast amounts of data and numerous potential failure points involved.

This scenario presents an ideal opportunity for generative AI, and today, AIOps startup BigPanda unveiled a new generative AI tool called Biggy, designed to expedite problem resolution. Biggy analyzes a diverse range of IT-related data, learning how a company operates, comparing it to ongoing issues and similar past scenarios, and suggesting solutions accordingly.

Since its inception, BigPanda has integrated AI into its systems, intentionally developing two distinct layers: one for data management and another for AI processing. This structure has set the stage for their transition to generative AI utilizing large language models. “Before generative AI, our AI engine was focused on developing various types of AI, all fed by the same data engine that powers Biggy and our generative, conversational AI efforts,” remarked BigPanda CEO Assaf Resnick.

Similar to other generative AI platforms, Biggy provides a prompt box for user queries and chatbot interactions. The underlying models are trained on data specific to the client organization and publicly available information related to various hardware and software, enabling them to tackle the types of challenges IT professionals encounter daily.

“The available LLMs are trained on extensive data sets and excel as generalists across all operational domains we address—be it infrastructure, networks, or application development. They also have a strong understanding of hardware nuances,” explained Jason Walker, Chief Innovation Officer at BigPanda. “For instance, if you ask about a specific HP blade server with an error code, it can effectively piece that information together, aiding significantly in processing event traffic.” However, merely looking up answers isn’t sufficient; the tool needs to provide deeper insight.

Biggy enhances its responses by synthesizing a variety of data types. “BigPanda ingests operational and contextual data from observability tools, change logs, Configuration Data Management Database (CDMB) files, and network topology. It also incorporates historical data and institutional knowledge, normalizing this information into key-value pairs or tags,” Walker added. While this may sound complex, it fundamentally means that Biggy evaluates system-level data, organizational context, and human interactions to deliver actionable insights.

When users provide input, Biggy analyzes all relevant data to generate an answer, ideally guiding engineers in resolving issues. While it acknowledges that generative AI isn’t infallible and outputs may vary in reliability, the platform informs users when there’s a lower certainty of accuracy in the response.

“For instances where we lack high confidence, we communicate that this is our best estimate and recommend human review,” Resnick noted. In cases with higher confidence levels, automation may be deployed, leveraging tools like Red Hat Ansible to address issues without manual intervention.

The process of data ingestion can be complex for clients, but it represents a critical step towards creating an AI assistant capable of diagnosing IT problems and facilitating quicker resolutions. While no AI is perfect, implementing an interactive AI tool promises to enhance current, often labor-intensive methods of troubleshooting IT systems.

BigPanda recently garnered attention after securing $190 million in new funding for expansion.