MCP Enterprise Use Cases: 7 Ways to Query Your IT Landscape

Enterprise architecture maps are often seen as repositories that get updated but rarely consulted. Too many clicks, too many filters, too much time to find an answer to what is often a straightforward question. The Model Context Protocol (MCP) changes this equation: with a connector between Claude and your UrbaHive architecture map, you ask questions in plain language and get a contextualised answer in seconds. Here are seven concrete use cases, drawn from real situations encountered by CIOs, architects, and security teams at mid-sized organisations.

What is MCP in this context?

A quick recap before diving into the use cases. MCP (Model Context Protocol) is an open standard that allows an AI assistant like Claude to connect to an external data source — in this case, your UrbaHive architecture map — and query it in real time. The map becomes a context source: Claude reads your data, reasons over it, and answers you. No data is modified; the connector is read-only. For a more technical overview, see our complete MCP guide and our article What is the Model Context Protocol?.

Use case 1: impact analysis before a system change

The situation. Before migrating middleware, updating an ERP, or decommissioning a server, the architect must identify every application and process that depends on it. In a complex architecture map, this analysis can take half a day.

With MCP. The question "Which applications are connected to the SAP PROD-01 server, and which business processes go through those applications?" delivers a structured list of direct and indirect dependencies in seconds. The architect can then prioritise regression tests and estimate the impact on the affected business units.

Value added. Impact analysis time reduced from hours to minutes. Lower risk of missing a critical dependency.

Use case 2: IT landscape summary for an executive committee

The situation. The CIO needs to present an IT landscape overview to the executive team or the board. Synthesising dozens of applications, servers, and flows into clear slides takes time and a cross-functional perspective.

With MCP. "Summarise our IT landscape in five points, mentioning the main functional domains, the number of applications per domain, and the critical infrastructure." Claude scans the architecture map and produces a structured summary that the CIO can use directly in a presentation or refine further.

Value added. Executive presentation prepared in minutes rather than hours. The summary is grounded in real architecture data, not approximations.

Use case 3: identifying dependencies of a critical server

The situation. A server shows signs of imminent failure or needs urgent replacement. The infrastructure team needs to know, in real time, which applications and data flows rely on that server.

With MCP. "Which applications hosted on the DB-FINANCE-02 server are running in production? What inbound and outbound data flows are documented for this server?" The answer arrives immediately, without having to dig through the architecture map or page the on-call architect.

Value added. Faster response during an incident. Better scoping of the impact perimeter within the first minutes of crisis management.

Use case 4: identifying business processes that cross a given application

The situation. The business team is considering replacing a CRM or document management tool. Before launching an RFP, they want to understand which company processes depend on that tool — to ensure the replacement covers all use cases.

With MCP. "Which business processes documented in UrbaHive go through the Salesforce application? For each process, indicate the functional domain and the services involved." Claude lists the identified processes, giving the project team solid input for the new tool's requirements document.

Value added. Quick functional mapping to prepare a replacement project or migration. Lower risk of missing secondary processes during the transition. See also our article on enterprise architecture and IT urbanisation.

Use case 5: onboarding a new developer or enterprise architect

The situation. A new team member — developer, enterprise architect, or project manager — joins the organisation. Understanding the IT landscape in a few days is a challenge: documentation is scattered, internal experts are busy.

With MCP. The new team member can ask Claude directly: "Explain our e-commerce IT architecture: which applications are involved, how they communicate, and what are the dependencies with the core IT system?" Claude answers based on the real architecture map, without anyone needing to organise a knowledge transfer session.

Value added. Shorter onboarding time. Internal experts are interrupted less for context questions. The new team member becomes autonomous faster.

Use case 6: preparing a NIS2 or DORA audit

The situation. The organisation must demonstrate to an auditor or regulatory authority that it controls its IT perimeter and critical dependencies. This requires producing inventories, flow maps, and lists of essential systems.

With MCP. "List all applications classified as critical in our architecture map, with their server dependencies and the business processes they support. Identify those that do not have a documented continuity plan." Claude produces a structured inventory that the security team can use as a working basis for the audit.

Value added. Faster preparation for regulatory audits. The UrbaHive architecture map becomes an active compliance tool, not a static document. Our articles on NIS2 and DORA detail the corresponding regulatory requirements.

Use case 7: planning a cloud migration

The situation. The organisation is considering migrating all or part of its IT landscape to the cloud. To build a realistic roadmap, the CTO must identify migration candidates, their dependencies, and associated technical constraints.

With MCP. "Which applications in our architecture map are hosted on our on-premise infrastructure? For each application, list its dependencies with other applications and any documented hosting constraints. Which applications would be priority candidates for a cloud migration?" Claude produces a preliminary analysis that the team can refine, without having to manually extract and cross-reference data across multiple tools.

Value added. A structured starting point for a migration project. Lower risk of migrating an application without identifying all its dependencies. For more on this, see our article on IT mapping and cybersecurity.

How to activate the MCP connector on UrbaHive

To benefit from these use cases, setting up the MCP connector takes less than ten minutes:

Log into your UrbaHive account and go to Settings > AI Connectors.
Generate a Personal Access Token (PAT).
Copy the provided JSON configuration and paste it into Claude Desktop or claude.ai.
Ask your first question.

The Starter (€29/month) and Professional (€99/month) plans include MCP connector access. See the Connectors page for details of capabilities available per plan.

Conclusion

MCP transforms the IT architecture map from a repository consulted by a handful of experts into a tool that every team can query. Impact analysis, audit preparation, onboarding, migration planning: the use cases are concrete and cover the core needs of a mid-sized organisation's IT department. The prerequisite is having an up-to-date, structured architecture map — and that is the foundational work UrbaHive is designed to support.

Create your free UrbaHive account and activate the MCP connector to query your IT landscape today.

FAQ

Does the architecture map need to be complete to use MCP?

No. MCP answers based on what is documented in UrbaHive. A partial map gives partial but useful answers. Gaps become visible during queries, which can motivate teams to enrich the documentation over time.

Are Claude's answers always reliable?

Claude relies on your architecture map data. If a dependency is not documented, Claude cannot know about it. Answers reflect the state of your map — which reinforces the value of keeping it current.

Can multiple users use MCP simultaneously?

Yes. The MCP token is linked to the organisation, not to an individual user. Multiple team members can use the connector in parallel, each from their own Claude client.

Is the MCP connector available on the Free plan?

The Free plan allows you to test the MCP connector with limitations (number of calls, features). Starter and Professional plans offer full access.

Can we create custom prompts for recurring use cases?

Yes. Claude Desktop and claude.ai allow you to save frequently used prompts. The team can build a library of standard questions tailored to their internal processes.

Internal links: