AI adoption & integration

Connecting legacy systems to AI

Your data is spread across ERP, Excel, shared folders and each department’s systems, where AI can’t see it. MCP connects them, so AI can query them directly within the permissions you set.

01

Who it’s for

  • You run legacy systems such as ERP, inventory software or an in-house database, and data lives in several places
  • You want staff to query internal data in plain language instead of asking IT for a report every time
  • You already use an AI assistant, but it only sees what someone pastes in by hand
  • Your legacy system has no API, or no one remembers how it works
  • ERP / inventory
  • Excel reports
  • Shared folders
  • In-house databases

MCP server

Read-only by default
Existing permissions kept
Every query logged

AI applications

Claude, Gemini…
Staff ask in plain language

The data stays in the systems it lives in. The MCP server opens only the queries you allow to AI, and logs every access.

02

What we do

Mapping data silos

We list where each system and dataset lives, who owns it, how it is accessed (API, database, files), and which data can be opened to AI.

Integration architecture

We decide which data is served through an MCP server, which AI application uses it, and how it connects: running locally or served over HTTP.

MCP server development

We build MCP servers for your legacy systems that provide query tools and data resources. They are read-only by default; any action that writes data gets its own confirmation step.

Permissions and audit

Existing accounts and permissions carry over, and every query is logged, so you can see who asked what through AI.

Security testing

We test the integration against LLM-specific risks such as prompt injection, to confirm the AI can’t be led by the data it reads into doing what it shouldn’t.

03

How it works

1. Map

List data sources and how they’re accessed, and pick the first system to connect.

2. Design

Define the queries and data to expose, the permissions, and how activity is logged.

3. Build and trial

Finish the MCP server for the first data source and have a small group of staff use it and give feedback.

4. Expand and hand over

Connect the next system based on how it’s used, and deliver documentation and a maintenance plan.

04

Deliverables

  • Data source inventory and integration architecture document
  • MCP server source code and deployment guide
  • Permission and audit log configuration
  • Security test report
  • Operations and maintenance documentation

05

Engagement cycle

Each cycle runs three months, six months or a year, depending on scope. At the end of each cycle we sit down with you and compare the results against the goals set at the start, then decide what the next cycle should cover, or whether to stop there.

Every cycle: audit → design → implement → check against the goals, then decide what's next

06

Pricing

Each engagement is estimated on its own: how many systems and how much data are involved, the people and time needed, and how long the cycle runs. Talk to us first and we’ll give you a number based on the actual scope, rather than quoting a price and then fitting the scope to it.

07

Common questions

What is MCP?
The Model Context Protocol, an open protocol that lets AI applications connect to outside data and tools in one standard way. Once a system has an MCP server, any AI application that supports MCP can use it.

Our legacy system has no API. Can it still be connected?
Usually, yes. An MCP server can read from a database, exported files or internal web pages. Without an API, we first work out the most stable way to read the data with the least impact on the original system.

Could the AI change our data?
By default it only runs read-only queries. Any action that writes data gets its own confirmation step and permissions, and is logged.

Will our data leave the company?
That depends on the AI application you choose and how it’s deployed. The MCP server can run inside your company; during design we show where each piece of data goes, and you decide.

08

Further reading

Tell us where you are

Email us about where you’re stuck, and we’ll reply with what could work and the next step.

Email us