Splunk launched Splunk AI Assistant, which will use generative AI and a domain specific large language model (LLM) built on its security and observability data.
The generative AI effort is part of Splunk AI, a portfolio of AI offerings announced at Splunk's .conf23 conference in Las Vegas.
According to Splunk, its AI Assistant will provide chat experiences to help users leverage Splunk Processing Language using natural language. Customers can use Splunk AI Assistant to write or explain custom queries. The big theme is that Splunk is looking to make its data more accessible to enterprises.
Constellation Research analyst Andy Thurai said:
"Splunk AI optimizes domain-specific large language models and machine learning algorithms specifically built on security and observability data. This means rather than having generic LLM models, their offering will have a domain specific knowledge trained on their data set. Splunk is helping customers train their own LLMs on their domain specific data, which can be powerful in unearthing tribal knowledge that is hidden in many corners of the enterprise.
Splunk AI Assistant and Splunk Processing Language (SPL) can be a good tool to help support personnel, SREs, SecOps folks and even DevOps teams, to find information faster for incident management whether it is security or service incidents."
While Splunk AI Assistant, which is in preview, was the headliner the company Splunk AI portfolio is aiming to free up various operations and engineering teams to do more strategic work and be more productive. Splunk said its AI efforts will be open and extend into its AI models as well as third party and proprietary tools.
Other Splunk AI and machine learning announcements include:
- Splunk App for Anomaly Detection provides AIOps related teams to automate anomaly detection and streamline operation workflows.
- IT Service Intelligence 4.17 improves detection accuracy with Outlier Exclusion for Adaptive Thresholding, which detects and omits outliers to provide more precise insights. ML Assisted Thresholding will use historical data and patterns to provide more accurate alerting with one click.
- Splunk Machine Learning Toolkit 5.4 offers guided access to machines learning technology. Security and IT operations teams can garner machine learning insights and bring external models into Splunk.
- Splunk App for Data Science and Deep Learning 5.1 will extend the Machine Learning Toolkit to bring together data science, machine learning and deep learning systems.
- The Splunk Threat Research Team has added 6 machine learning detections to Splunk Enterprise Security to keep up with new threats and attacks.
More:
