This article covers Kentik AI, including data privacy and security considerations.
The Kentik AI process flow for Journeys and Query Assistant.
About Kentik AI
Kentik AI, a comprehensive suite of AI capabilities, enhances network observability and operations.
Kentik AI Features
Kentik AI features includes:
Query Assistant: Translates natural language questions into Metrics Explorer queries for faster, more natural interaction with NMS metrics data.
Journeys: Provides an AI-assisted conversational user experience for troubleshooting and network analysis, translating natural language questions into queries executed in Data Explorer and Metrics Explorer. It’s contextually aware of previous queries, allowing for iterative querying and follow-up questions.
Cause Analysis: On-demand or automatically triggered analysis in Data Explorer that identifies the most contributing factors for sudden traffic changes, spikes, or drops.
AI Enhanced Insights: New AI-driven analysis and correlation from multiple data points and sources, providing valuable information to users in Kentik Insights.
Summaries: Simple breakdowns of complex network issues with solutions and tips, displayed in the Portal.
Note: Check your Kentik AI settings page (see Kentik AI Settings UI) for enabled features (requires Super Administrator privileges).
Kentik AI Models
Kentik AI uses GenAI Large Language Model (LLM) services from leading companies:
OpenAI models hosted by OpenAI.
Foundational models from Anthropic, Google, or others hosted by Google Cloud's Vertex AI service.
How Does Kentik Use AI?
Kentik AI is used in various contexts across the platform for specific and general capabilities.
Journeys and Query Assistant
The GenAI service translates user queries into a structured format for Kentik’s Query system:
Access via API: Kentik uses enterprise API access to the GenAI service provider, ensuring information sent to the LLM is not retained for training purposes (terms and conditions apply). For more information, see Kentik AI Privacy & Security.
No model training: Kentik uses sophisticated Prompt Engineering techniques, e.g., "few-shot prompting," to improve accuracy without training or fine-tuning the models.
Submitted data: The data submitted to the GenAI service includes the elements detailed in GenAI Input Data.
Note: The network metadata can be expanded in the future to improve user experience and develop new features.
Journeys query techniques: Journeys uses specific techniques outlined in Queries for Journeys to enable responses informed by prior prompts.
Prompt protection: Kentik takes precautions to guard against LLM prompt injection:
Check if a valid Kentik query comes back from the GenAI service.
Responses that don't fit the expected schema and can't be executed by the Kentik query engine are discarded.
LLM queries are executed in the user context, isolating them at the user’s company level.
Queries are read-only, reducing the risk of injection-type attacks.
GenAI Input Data
The data submitted in the input to the GenAI service contains the following elements:
User’s Question: A question typed in the Query Assistant or Journeys input prompt.
Query Context: The names of metrics, dimensions, and measurements used in the Kentik Data Explorer and NMS Metrics Explorer query schema.
Network Metadata: A set of relevant metadata about the user’s network, including:
Site names
Site’s countries
Site Markets
Provider names used in interface classification feature
Device Labels (”Labels” used in Kentik portal)
AS Groups
DDoS Alerting policy names
Aggregated Data Results: For Journeys, results returned from successive user prompts shape the next query (see Queries for Journeys).
Queries for Journeys
The system includes the aggregated table results of the previous query. Only the two most recent Top-X aggregated table results visible on the table are included in the LLM input message. This feature enriches user communication by allowing them to ask questions which are related to previous query results, such as filtering results based on a specific row or showing different metrics for network objects from the first query.
Insights, CA, and Summaries
The GenAI service interprets query or analysis results and makes them more user-friendly.
Kentik analyzes network telemetry data using algorithms, data mining, or machine learning techniques.
The data is usually related to a specific problem, such as network anomalies, metric thresholds, or public cloud connectivity.
The analysis results form an aggregated data structure containing relevant information for the problem.
This aggregated data is used as input for the GenAI service to provide a user-friendly explanation and enrich the user experience.
Kentik AI Privacy & Security
Kentik ensures that Kentik AI respects user privacy and security through the following practices.
Protection of Prompts
Kentik stores user questions submitted to Query Assistant and Journeys internally for troubleshooting and improving user experience, following our Security Overview.
Kentik’s input prompts to the provider’s GenAI service is handled according to their privacy policies (see Privacy & Security Resources), summarized as follows:
Submitted data is not used to train any GenAI model.
OpenAI can retain API inputs and outputs for up to 30 days to provide services and identify abuse. They are removed after 30 days from their systems unless further retention is legally required.
Google does not store API inputs or outputs.
Data is encrypted in transit.
The OpenAI API platform is SOC 2 compliant, while the Google Vertex AI platform is SOC 2, ISO 27001 and ISO 27701 compliant.
Customer Control of AI
Kentik customers retain control Kentik AI access at the company level. Users with Super Administrator privileges can enable/disable Kentik AI features on the portal's Kentik AI settings page, accessed via the Organization Settings menu (see Enable Kentik AI switch).
Enabling Kentik AI enables the use of Query Assistant, Journeys, and other AI-assisted features
Disabling Kentik AI disables the use of all Kentik AI related features.
Privacy & Security Commitment
Kentik proactively enhances security measures to ensure customer data integrity and confidentiality. For help with any questions on this topic, contact our Customer Success team.
Privacy & Security Resources
The following resources provide additional information about privacy and security policies related to Kentik AI:
Kentik: Security Overview (Kentik Knowledge Base)
Vertex AI: Google Cloud Vertex AI, Generative AI and data governance.
Kentik AI Legal Disclaimer
Kentik's AI features send user queries and contextual info to LLM providers, including OpenAI. We ensure these providers don’t use this data to train or refine their AI systems. However, we encourage customers to ensure any AI use aligns with their corporate policies. AI may contain mistakes and the use of these features is as-is without warranties. You must agree to these terms to use these features. For more on Kentik's approach to trust & safety, please see the linked documentation.
© 2014-25 Kentik