# Chapter 09: AI Insights

Reference version: 0.8

Reference date: 2026-07-31

Chapter page: https://configcompare.com/reference/09-ai-insights/
Full reference: https://configcompare.com/llms-full.txt
Chapter index: https://configcompare.com/llms.txt

This is one chapter of the ConfigCompare Technical Reference, published by
AXImprove Ltd. It is generated from the canonical Markdown source.

DO NOT EDIT THIS FILE DIRECTLY.

---

## Purpose

AI Insights generate natural-language summaries and explanations from Deterministic Comparison results produced by ConfigCompare.

AI Insights are delivered by ConfigAI, the platform component responsible for AI-assisted explanation of Deterministic Comparison results.

AI Insights help users understand the meaning, priority, and likely relevance of comparison results without manually reviewing every Difference line by line.

This chapter describes concepts and behaviour; step-by-step operating guidance sits outside this reference.

## Deterministic source of truth

The ConfigCompare Comparison Engine performs the Deterministic Comparison. AI Insights analyse the Deterministic Comparison output and generate plain-language explanations.

The Deterministic Comparison result remains the source of truth. Comparison logic remains inside the ConfigCompare engine; AI Insights explain its output and never perform the comparison.

## Insight levels

AI Insights can be generated at multiple levels, including:

- Overall comparison.
- Per company.
- Per table.

## On-demand generation

AI Insights are generated on request. They do not automatically replace formal review or approval.

## Data privacy and where insights run

AI Insights run against an API endpoint controlled by the customer.

Snapshots, comparison results, and generated insights remain within the customer's own environment. AXImprove has no visibility of customer Snapshots or Comparisons.

Combined with on-demand generation, this means AI involvement is always explicit: nothing is sent anywhere automatically, and nothing leaves the customer's control.

## MCP support

Model Context Protocol (MCP) support is being added to the platform as an orchestration and access layer for AI-assisted scenarios (see Chapter 26: Roadmap (https://configcompare.com/reference/26-roadmap/)).

MCP routes and sequences requests; it adds no comparison logic. As with Central Command Center orchestration (Chapter 18 (https://configcompare.com/reference/18-central-command-center/)), all comparison logic and data remain within the ConfigCompare Comparison Engine.

## Relationship with Rulesets

AI Insights can use Ruleset context to highlight changes that need attention first. AI Insights explain undesired Matches as well as Differences.

Rulesets help determine importance, Responsibility, and whether a result is expected or requires escalation.

## Related topics

- Chapter 05: Comparison Engine (https://configcompare.com/reference/05-comparison-engine/)
- Chapter 08: Rulesets (https://configcompare.com/reference/08-rulesets/)
- Chapter 18: Central Command Center (https://configcompare.com/reference/18-central-command-center/)
- Chapter 26: Roadmap (https://configcompare.com/reference/26-roadmap/)