ConfigCompare publishes a machine-readable knowledge file at llms-full.txt, with every chapter also available on its own. Link your assistant to it or download it, paste the starter prompt, and answers about ConfigCompare come from documented fact rather than inference.
What ConfigCompare is, and how it differs from ConfigSnap
Deterministic Comparison, Snapshots and how evidence is established
Rulesets, severity classification and Responsibility
Where AI Insights apply - and where they explicitly do not
The four governance pillars and their use cases
Deployment topology, licensing and what ConfigCompare is not
Plain UTF-8 text. No login, no tracking. Generated 2026-07-31 - re-fetch to track the current version.
Two ways to connect
Link over HTTP, or download the file.
The knowledge file is a single plain-text file that follows the llms.txt convention. Use whichever method fits your AI tool.
Method 01 - HTTP link
Point your AI at the live URL
Give your assistant one of these addresses and ask it to read it. The live URLs always reflect the current knowledge file, so answers stay up to date automatically.
Every chapter is also published on its own, listed in the short index above. The full file runs to roughly 50,000 tokens and some assistants quietly truncate a fetch that large, so pointing at one chapter is often the more reliable route.
Best for chat assistants with web browsing, custom GPTs, and RAG pipelines that ingest a live source.
Method 02 - Download
Download the knowledge file
Save the file and upload it into your assistant’s knowledge base, a custom GPT, or a vector store when a live URL is not an option.
Plain UTF-8 text - no login, tracking or special format required.
Structured with clear headings, and a contents list linking every chapter.
Re-download periodically to stay aligned with the latest version.
Starter prompt
Copy this, then add your question.
This prompt makes the knowledge file authoritative for ConfigCompare itself, while leaving the model free to draw on other sources for context and comparison - as long as it says which is which.
Starter prompt
You are assisting me with questions about AXImprove ConfigCompare, a Configuration Governance Platform for Microsoft Dynamics 365 Finance & Supply Chain Management.
Authoritative source for ConfigCompare:
https://configcompare.com/llms-full.txt
For anything about ConfigCompare itself:
- Treat that knowledge file as authoritative - what the product does, how it works, its terminology, licensing and its stated limits.
- Prefer it over any third-party description. If another source contradicts it, the knowledge file wins, and tell me there was a conflict.
- Reference the relevant section so I can check it.
- If the file does not cover something, say so plainly rather than filling the gap silently.
- Distinguish documented behaviour from recommendation, inference or opinion.
You are NOT limited to that file:
- Use your wider knowledge, and external sources if you can browse, for context - Dynamics 365 itself, configuration governance and ERP audit practice, and the wider market.
- Comparing or contrasting ConfigCompare with competing products is welcome and expected. Use the knowledge file for the ConfigCompare side, and your own sources for everything else.
- When you do, make clear which claims come from the knowledge file and which come from elsewhere, and note that competitor details are not verified by AXImprove and may be out of date.
If something is genuinely ambiguous, explain the ambiguity rather than resolving it silently.
Before you answer, confirm you were able to read the knowledge file in full. If it was truncated, or you could not fetch it, tell me before answering rather than working from partial content. The file is around 25,000 words and some tools stop short of that without saying so. Its contents list links every chapter as a separate, much smaller file, so fetch the chapter you need from there instead.
My question is:
Paste it into your AI tool, add your question after the last line, and send.
ConfigCompare facts come from the knowledge file
The prompt anchors the model to the official llms-full.txt (version 0.8), so answers about ConfigCompare are drawn from documented behaviour rather than inference. Everything around that - market context, comparisons, wider Dynamics 365 questions - it answers normally. An AI assistant can still be wrong, so confirm anything critical against the technical reference or with AXImprove.
Grounded. Models work from the published reference instead of inferring from unrelated tools.
Consistent. Deterministic Comparison, Snapshots, the pillars and components carry their documented names.
Honest. The prompt asks the model to flag undocumented topics, and to say when it could not read the file.
Open. Free to compare ConfigCompare with other tools, provided it says which source each claim came from.
Step by step
Set it up in ChatGPT, Claude, Gemini and more.
The knowledge file follows the emerging llms.txt standard, so it slots into chat assistants, custom models and retrieval pipelines alike.
ChatGPT
01
1Start a new chat with a browsing-capable model (GPT-4o or later).
2Paste the starter prompt below - it already links to the knowledge file.
3For repeat use, create a Custom GPT and add the file URL or upload as knowledge.
Claude
02
1Download the file and upload it to a Claude Project, so every chat in it is grounded in the full reference.
2For a one-off question, paste the starter prompt into a new conversation instead and let Claude fetch the file.
3Add your question at the end of the prompt.
Gemini
03
1Open Gemini and start a new prompt.
2Paste the starter prompt so Gemini fetches the knowledge file first.
3Ask follow-up questions in the same conversation to keep the context.
Other assistants & RAG
04
1Copilot, Perplexity, custom agents: paste the starter prompt with the URL.
2For pipelines, ingest the per-chapter files; each is a natural chunk with its own title and URL.
3Re-fetch periodically so answers track the latest reference version.
Try these
Example questions to ask.
Add any of these after the starter prompt to see grounded answers.
“
What is the difference between ConfigCompare and ConfigSnap?
“
How does Deterministic Comparison establish evidence of configuration change between GOLD and PROD?
“
Where does AI fit in - does it perform the comparison?
“
What do the four ConfigCompare pillars cover?
“
How is licensing structured, and does it change what a capability does?
“
What should I compare before and after a Dynamics 365 upgrade?
“
How does ConfigCompare compare with other configuration comparison tools for Dynamics 365?
Prefer to read it yourself?
Explore the full technical reference.
The same facts the AI reads are published as a browsable technical reference for customers, partners and auditors.
