Natural language, Supernatural results

Sumo Logic Mo Copilot is an AI assistant that is integrated into the Sumo Logic Log Analytics Platform. It combines a contextual experience with natural language queries to help users quickly drive relevant insights from logs. Copilot does not process your logs, and none of your data is shared with any additional third party. Rather, Copilot enables you to troubleshoot and investigate incidents using context inferred from your logs.

Yes. You can use Copilot to search and extract relevant information from unstructured logs, ensuring no critical insights are missed during investigations. Field extraction rules are required. Semi-structured (JSON + unstructured payload in a log message) should also work.

Yes. Copilot retains conversation and search history, allowing users to resume investigations where they left off, maintaining context and continuity.

Copilot uses AI to interpret natural language queries and recommend relevant search results and query refinements, making it easier for users to find key insights quickly.

All of Sumo Logic’s machine learning (ML) features undergo legal, compliance and security reviews to ensure they serve customer outcomes, data minimization, fit-for-purpose data and anonymization.

In Sumo Logic Mo Copilot, the schema of logs and sampling of field values are provided as context to an AI. Field values can contain PII or confidential data. For example, email or IP addresses are PII and often, confidential data as well. However, to be useful, Copilot has to enable insights about such data.

No. No customer data or PII is used for training or other purposes. All our capabilities serve customer outcomes. Our classic ML capabilities (e.g. AI-driven alerts and its anomaly detection features) create customer-specific models. Sumo Logic Mo Copilot uses a Large Language Model (LLM) served via Amazon Bedrock. As explained in our documentation and included links, no customer data is used for training or other purposes in the case of Sumo Logic Copilot.

Some of our classical ML models store customer data in our ML pipelines to optimize performance. For example, our AI-driven alerts feature log anomaly detection and build ML models from 60 days of logs. To accomplish this, we retrain the model once a week. In this example, each week, we add one week of new data while expiring the oldest week of data. Rolling data windows are done to avoid fetching 60 days of data for every training run.

Sumo Logic Copilot also stores customer data in the ML backend to optimize performance. For example, certain Copilot features rely on the history of a customer's queries. We will expire such data on a rolling window basis.

Yes. To opt out of Sumo Logic Copilot, a support ticket is required.

For Generative AI, we use a foundation model served by Amazon Bedrock as explained in our documentation. Our classical ML features use various open-source Python libraries approved by Sumo Logic.

Sumo Logic Copilot is an ensemble of Generative AI (GenAI) and classical ML techniques. Other ML capabilities, such as AI-driven alerts, typically use an ensemble of classical ML approaches.

Yes, the on-call developer or security engineer troubleshooting an incident is the expected user.

The on-call developer or security engineer troubleshooting an incident is the expected user. They interact with Copilot using Natural Language questions or through contextual suggestions.

No. The foundation model provider used by Amazon Bedrock has no access to customer data.

No.

For Sumo Logic Copilot, the GenAI model is not accessible to Sumo Logic as it is licensed through Amazon Bedrock.

We launched our first release of Copilot on Dec 2, 2024. Every major capability added to Copilot (and AI/ML in general) goes through legal, compliance and application security reviews. These reviews typically coincide with releases that offer new insights or analytics and process data that was previously not used.

No, because the GenAI model used in Copilot is not accessible to Sumo Logic. For components in our control, we follow industry best practices for code reviews and change management. On-going monitoring and troubleshooting of our AI/ML features use logs telemetry analyzed by Sumo Logic's Log Analytics Platform.