AI Recommendations: Ask Anything About Your Data for Instant Guidance

Our AI taps into your tools and past incidents to provide real-time, expert guidance and instant answers, empowering on-call engineers to resolve incidents faster.

Jessica Abelson, Director of Product Marketing
Jul 10th, 2024

Picture this: Dave, a newly hired on-call engineer, gets a buzz on his phone — Nginx 4xx status codes are spiking. Still getting accustomed to his on-call duties and unfamiliar with Kubernetes or Nginx, Dave isn’t sure where to start or whom to ask for help.

In the past, Dave might have spent precious time sifting through documentation, reaching out to senior engineers, and piecing together a solution. But today, he has a powerful tool at his disposal: his AI Teammate.

Dave opens Slack and types a simple question to his AI Teammate: “We are seeing a ton of 4xx errors in our Nginx logs. What are the first 3 things we could do to triage the issue?” Immediately, the AI Teammate springs into action.

Drawing on its deep integration with Dave’s environment, it gathers knowledge from across his disparate tools and systems to offer precise suggestions and recommend actionable next steps.

See it in action:

Transforming Unstructured into Actionable Insights

Data isn’t the problem — teams are practically swimming in it. The issue is finding and using the right data when the clock is ticking. Roughly 90% of operational data is unstructured. The answers that on-call engineers need usually exist but are buried in an avalanche of tickets, documents, and Slack conversations. Observability tools help make sense of structured data (logs, graphs, metrics), but institutional knowledge — the tried-and-true learnings gained through experience — is a hazy mess.

Our AI Teammate is an expert at sifting through your unstructured human data and content to unlock those learnings and insights, and offer them to you at precisely the right moment.

Step-by-Step Guidance: How Our AI Teammate Guides On-Call Engineers from Alert to Resolution

AI Knowledge Base

In the above scenario, our AI Teammate begins by identifying the potential causes of the issue, leveraging data from historical incidents and your tools (i.e. Jira tickets, Slack conversations, Confluence docs) to correlate patterns and provide vetted next steps. It pulls from a runbook that already exists in Dave’s environment — suggesting the steps he should take to analyze the Nginx logs.

It recommends inspecting the access logs for unusual activity, identifying the top endpoints generating the errors, and checking the configuration files for any syntax errors, providing the exact query for each step.

For Dave, this is invaluable. He doesn’t have to guess or blindly follow generic troubleshooting steps. Each recommendation is tailored to the specific context of the incident, informed by past occurrences and expert knowledge.

And if Dave still needs help? Simple. Our AI leverages contextual information from Slack, Confluence, Jira, and GitHub to dynamically identify the right experts for the incident.

Proactive Suggestions

As Dave works through the steps, our AI monitors the Slack conversation, detecting the progression of the incident. When it notices changes, it suggests updating the incident status and even writes the status update for him.

Dave can also easily generate an incident summary and update using AI, ensuring the entire team is kept in the loop with little effort.

Intelligent Event Logging

Our AI Teammate is continuously logging key events — data intelligently accumulated from written data (e.g. chats, documents, tickets), actions taken (e.g. running a script to roll back a change), and voice data (Zoom, Teams, or Google Meets meetings). By doing so, it gets better over time, meticulously gathering knowledge that can be used during future incidents.

AI-Driven Automation

Our AI Teammate goes beyond offering recommendations — it actually helps you take action faster by surfacing the appropriate automation for a given scenario.

Imagine Dave encountering a recurring issue that requires a series of manual steps to mitigate. Instead of spending time executing each step, our AI Teammate identifies the appropriate automation to handle the task. For example, it might suggest running a script to restart a service or roll back a recent change, significantly reducing the time and effort required from Dave.

Just as our automation philosophy has always kept humans in the loop, our AI ensures humans are always in the driver's seat, able to make the final call on any action. This approach maintains a balance between leveraging AI's efficiency and preserving human oversight, ensuring that engineers have full control over their environment.

A 24/7 Incident Assistant

Designed to be a 24/7 on-call assistant, our AI Teammate empowers on-call engineers to expertly manage any incident — whether they’re in their first month or their tenth year. In Dave’s case, he’s able to resolve the Nginx issue faster and more effectively. What could have been a stressful and drawn-out process becomes a manageable task, empowering Dave to confidently handle on-call responsibilities, even as a newcomer.