Splunk AI: Ask me Anything - 3/20/25

Community Office Hours

Splunk AI: Ask me Anything - 3/20/25

1 Comment
Cover Images - Office Hours (11) copy.png
Published on ‎01-15-2025 01:28 PM by Splunk Employee | Updated on ‎03-31-2025 11:58 AM

Register Here. This thread is for the Community Office Hours session on Splunk AI: Ask me Anything on Thurs, March 20, 2025 at 1pm PT / 4pm ET

 

Ask the experts at Community Office Hours! An ongoing series where technical Splunk experts answer questions and provide how-to guidance on various Splunk product and use case topics.

 

What can I ask in this AMA?

  • How does Splunk use generative AI?
  • What are some common use cases to get started with the Machine Learning Toolkit (MLTK) app?
  • How can Splunk help me monitor and secure my generative AI applications?
  • How do the Splunk AI Assistants keep customer data confidential?
  • How can I leverage ML for anomaly detection?
  • Can I develop and package a custom machine learning model in Splunk?
  • Anything else you’d like to learn!

 

Please submit your questions at registration. You can also head to the #office-hours user Slack channel to ask questions (request access here)

 

Pre-submitted questions will be prioritized. After that, we will open the floor up to live Q&A with meeting participants.

 

Look forward to connecting!



Labels (2)
0 Karma
adepp
Splunk Employee

Hi everyone! Here are a few questions from the session (get the full Q&A deck and live recording in the #office-hours Slack channel):

 

Q1: What is new in MLTK 5.6?

  • LLM Integrations: New feature to allow customers to connect to external LLM models, from providers including OpenAI, Gemini, AWS Bedrock, Anthropic, Grok, Microsoft hosted OpenAI and Ollama
  • Enhanced ONNX apply: ONNX apply now supports multi-variate outputs
  • Upgrading OpenSSL to v3.4 to mitigate a critical vulnerability in PSC 4.2.3 and 3.2.3

Q2: Can I use indexed data to train the AI, if so how?

  • Training models directly in MLTK/DSDL
    • Can index data and prepare it with SPL for fit/apply commands to implement on ML workload
    • Key to utilize summary index for better performance
  • Using genAI via MLTK
    • Search command sends data to external LLM
  • Vector databases & Splunk
    • Setup a LLM/RAG pattern to contextually pass documents from vector DB
    • Encode Data from Splunk (logs, documents, searches, etc.)

Q3: What are some common use cases to get started with the Machine Learning Toolkit (MLTK) app?

Other Questions (check the #office-hours Slack channel for responses):

  • What’s Splunk doing around AI?
  • What new in Splunk Data Science and Deep Learning 5.2?
  • How does Splunk manage data privacy with the AI Assistant? What data is shared and how is it used?
  • Is there a cost to use the AI assistant? And how does it impact my SVC utilization?
  • Is the AI Assistant for SPL conversational?
  • How can Splunk help me monitor and secure my generative AI applications?
  • Is it possible to enable AI Assistant in Observability Cloud on certain services or certain roles/groups?