---
title: "March 2026 - AI / Insights"
slug: "march-2026-ai-insights"
updated: 2026-06-29T20:41:23Z
published: 2026-06-29T20:41:23Z
canonical: "kb.kentik.com/march-2026-ai-insights"
---

> ## Documentation Index
> Fetch the complete documentation index at: https://kb.kentik.com/llms.txt
> Use this file to discover all available pages before exploring further.

# March 2026 - AI / Insights

The following updates were released to the Kentik platform for AI / Insights during March 2026.

## Features

- **AI Auto Investigations**: New Auto Investigations tab in Kentik AI lets you create investigation configurations tied to specific policies, enable/disable automatic investigations per policy, set an investigation delay, and assign notification channels and runbooks.
- **AI Advisor report subscriptions**: Schedule AI Advisor conversations to run automatically on a recurring basis and receive the results by email with a PDF attachment — ideal for recurring NOC/shift summaries.
- **AI-generated notification summaries**: Email and Slack alert notifications now include AI-generated short and long summaries of network issues, with improved formatting and severity badges.
- **AI Advisor token budget management**: Conversations handle large contexts more gracefully through soft/hard token limits and progressive message compression.
- **Smarter model lifecycle**: Deprecated AI models are transparently redirected to their replacements at request time, so existing integrations keep working without manual changes.

## Bug Fixes

- **KMI export filenames**: Market Intelligence exports now include a unique timestamp (and the ASN number on ASN exports) to prevent filename collisions across multiple exports.
- **AI Advisor PDF export robustness**: Link resolution now correctly handles code blocks and inline code so exported PDFs don't rewrite literal code examples.
- **AI Advisor wide table scrolling**: Tables with many columns now scroll horizontally instead of overflowing.
- **AI Advisor emoji rendering**: Fixes a case where emoji failed to render in AI Advisor markdown responses.
- **AI Advisor inline code styling**: Inline code now uses a subtle background that integrates better with surrounding text.
- **Rich entity popovers**: Improves reliability and correctness of entity popovers, including special-character handling and a fix for 0 Mbps display.