Introduction

Two pillars that feed your AI the materials to judge

IceShore has two pillars: a web service where you review and correct, and an MCP (the standard that passes data to your AI) that returns the result. The foundation is the ADL (Architecture Description Language) — the language that describes business meaning.

Two pillars

In the web app, IceShore displays the architecture (ADL) your AI has built — showing you how your AI reads your business, so you can refine it. MCP then returns the refined data to your AI.

MCP Returns the data to your AI Passes the ADL

Your AI business understanding

Web service (app) Review & correct Visualize business meaning via ADL

The Web service (app) lets you review & correct the business meaning; the MCP delivers that corrected meaning (the ADL) to your AI.

A use-case-centric ADL and its interpretation engine

Starting from the use cases — who uses the system and why — IceShore organizes the affected user scale, processing flows, and downstream impact into the decision materials your AI needs to assess the impact of a change.

START Use-case-centric ADL business meaning
Interpretation engine interprets the ADL
Decision materials (3 axes + traffic)
- Business (user scale)
- Processing (flows)
- Downstream (ripple)
- + traffic
Your AI assesses the impact

It starts from the use cases (business meaning). The interpretation engine prepares the materials, and your own AI uses them to assess the impact. How well these decision materials line up across the three axes — business, processing and downstream impact — shows as AI understanding (Good / Needs Review / Needs Action). IceShore prepares the materials; your AI does the judging and answering.

Connect your AI → See what you can do