AI assistants like Claude, ChatGPT and Microsoft Copilot are remarkably good at answering questions, summarising information and drafting reports. But ask one how your business is tracking against budget and it hits a wall: it simply can’t see your numbers. That’s the gap an MCP is designed to close. In this article we explain what an MCP is in plain English, how it works, and what Jedox MCP means for finance teams who want to use an external AI with their live planning data.
In the video below, Stephen Hambling, founder and CEO of Kybos, explains what an MCP is and why it matters for finance teams.
What is an MCP?
MCP stands for Model Context Protocol. It’s an open standard that lets AI tools connect directly and securely to the business systems where your data lives – your planning model, ERP, CRM or data warehouse.
The simplest way to think about it is as a universal adapter between AI and your data. Before MCP, every AI tool needed its own custom integration with every system it wanted to talk to. MCP replaces that tangle with one common language: if a system offers an MCP connection, any MCP-compatible AI tool can use it.
MCP was introduced by Anthropic, the company behind Claude, in November 2024 and was quickly adopted across the industry. In December 2025 it was donated to the Linux Foundation’s Agentic AI Foundation, so it’s now governed as a neutral, open standard rather than owned by any single AI vendor. For finance leaders, that matters: you’re not betting on one AI provider when you adopt it.
The problem MCP solves for finance teams
Without a connection like MCP, most finance teams use AI in the same way: export the data to a spreadsheet, paste it into a chat window, ask the question and hope the figures are still current. It works, after a fashion, but it creates three problems.
- It’s slow. Every question starts with an export, a tidy-up and a paste – the same manual effort that keeps teams on the Excel hamster wheel.
- It’s out of date instantly. The moment you export, the data is a snapshot. Any reforecast or late journal leaves the AI working from the wrong numbers.
- It’s uncontrolled. Sensitive financial data ends up copied into files and chat histories outside your governed systems, with no clear view of who has seen what.
MCP removes the copy-and-paste step entirely. The AI asks your system for the data it needs, when it needs it, through a controlled connection.
How does an MCP work?
You don’t need to be technical to understand the basics. There are three parts:
- The AI tool – for example Claude – where you type your question in plain English.
- The MCP server – a connector provided by the business system (in this case Jedox). It tells the AI what it’s allowed to do, such as “read this report”, “look up these KPIs” or “create a new dimension”.
- Your business system – the Jedox planning model holding your live budgets, forecasts and actuals.
When you ask a question, the AI works out what information it needs, requests it through the MCP connection, receives live figures back from your system and then explains the answer. For example, you might ask: “Why is EMEA revenue behind forecast?” Claude asks Jedox through MCP, Jedox returns the relevant figures from your planning model, and Claude explains what’s driving the gap – all without anyone running an export.
What is Jedox MCP?
Jedox MCP is Jedox’s connector for external AI. It gives AI tools such as Claude, ChatGPT, Google Gemini and Microsoft Copilot – as well as custom AI agents – governed access to your Jedox planning model and reports. It sits alongside the built-in capabilities of JedoxAI and the wider AI features introduced in the Jedox 26 major release, extending them to whichever AI tool your organisation has chosen to standardise on.
In practice, Jedox MCP lets finance teams do three things.
1. Ask questions of live data
Ask questions in plain English and get answers grounded in your live planning model rather than a static export – from retrieving KPIs to explaining variances.
2. Build and extend your model
Describe what you need and let AI do the groundwork: create a new dimension, extend a hierarchy, write planning logic or build a report. You can even drop in a raw data file, describe what you want, and have AI propose a model structure for you to review.
3. Stay in control
Where AI takes actions, runs workflows or writes data back, those changes follow your approval processes, with people firmly in the loop.
Is it secure? Why governance matters
For any CFO, the first question about connecting AI to financial data should be about control. Jedox MCP is designed around the security you already have:
- Same roles and permissions. The AI only sees the planning structures and data that the user is already authorised to see in Jedox. If you can’t see payroll, neither can your AI assistant.
- No uncontrolled copies. Data is accessed through the connection rather than exported into spreadsheets and chat windows.
- A clear audit trail. Transparency and auditability help you maintain the lineage behind decisions.
There is one important caveat. AI is only as good as the data it can reach. As Stephen Hambling explains in Stop playing AI Jenga with your FP&A, connecting AI to a weak foundation simply gives you faster wrong answers. A well-structured planning model and good-quality data are what make MCP genuinely valuable.
What this means for CFOs and FP&A teams
MCP changes AI from something you feed data into, to something that works directly with your planning model. The benefit isn’t just convenience – it’s time. Less effort spent gathering and reconciling numbers means more time spent on the analysis and decisions that matter, which is why we argue that time is the number one benefit of AI in EPM and FP&A.
It also future-proofs your AI strategy. Because MCP is an open standard, your Jedox model isn’t tied to one AI vendor. If your organisation moves from one AI tool to another, the connection to your planning data stays the same.
If you’d like to understand more about where AI fits in FP&A, our articles on the 3 types of AI changing FP&A forever and how FP&A can take advantage of AI are a good place to start.
How to get started with Jedox MCP
Jedox MCP is available as an add-on to the Jedox Full Suite user licence, and includes a Jedox Claude Plugin with best-practice cookbooks to help you get value quickly. Setup varies depending on your chosen AI tool and your Jedox environment, which is where an experienced partner helps.
To see it in action, join our live webinar Think it, build it, report it with Jedox AI, where we build a live model from a raw file using Jedox MCP and Claude. Alternatively, arrange a live demonstration or schedule a call to talk through how Jedox MCP could work for your finance team.
Frequently asked questions
About Kybos
Kybos is a dedicated UK & Ireland Jedox Platinum partner. We build planning and analysis solutions that deliver value fast using accountancy qualified consultants. Whether you want a fully customised application or to build upon an existing solution, Kybos consultants are here to help.