Every software vendor now says it has AI. The more useful question for a finance leader is simpler: what problem does it actually solve? This article answers that for JedoxAI, the AI built into the Jedox planning platform. Rather than list features, we start with the problems finance teams tell us about every week, and show how JedoxAI deals with each one.
What is JedoxAI?
JedoxAI is the AI layer inside Jedox. It combines three types of AI: machine learning for predictive forecasting, large language models so you can ask questions in plain English, and agentic AI, where specialist “agents” carry out tasks for you. If those terms are new, our Diary of a CFO episode on the 3 types of AI changing FP&A explains them simply.
Because JedoxAI sits on top of the same model your budgets, forecasts and reports already live in, it works with your real numbers, your structure and your permissions. That is what separates it from a general chatbot.
Problem 1: getting a simple answer takes too long
A director asks for sales by product for the last four years, actuals only. It is a reasonable question, but it usually means a request to finance, someone rebuilding a report, and an answer a day or two later.
How JedoxAI solves it: with the Reporting Agent you type the question as you would say it, for example “show me only actuals for 2022 to 2025 in the columns”, and the view updates in seconds. No formulas, no pivot tables and no waiting for an analyst. Anyone with access can get their own answers, and finance gets its time back.
Problem 2: you know what happened, but not why
Most management packs are good at showing the “what”: revenue is down 6% against budget. Explaining the “why” is where the hours go, digging through cost centres, products and regions to find what actually moved.
How JedoxAI solves it: Data Insights, introduced in Jedox 26, analyses KPIs and reports to pinpoint the key drivers and explain performance in plain business language. A one-click “summarise information” gives you a first draft of the story. You start your variance analysis with the explanation rather than the hunt.
Problem 3: forecasts built on “last year plus 5%”
Many forecasts are still last year’s numbers with a percentage added. It is quick, but it ignores seasonality, trends and the signals hidden in your own data, and it is hard to defend when the board asks where a number came from.
How JedoxAI solves it: machine learning learns from your historical data to produce forecasts for revenue, cash flow, demand and churn. It tests several forecasting methods, ranks them by accuracy and shows the drivers behind the result, so the forecast is explainable rather than a black box. It gives you a strong, objective baseline that your team can then adjust with their own knowledge. It also fits naturally with rolling forecasts and driver-based planning.
Problem 4: problems are spotted too late
An unusual cost, a sudden drop in orders from one customer or a data load that went wrong often isn’t noticed until month-end, or until the board pack is already out.
How JedoxAI solves it: anomaly detection flags numbers that don’t fit the expected pattern, so you can investigate while there is still time to act. Churn prediction does the same for customers, highlighting those most at risk of leaving before they go.
Problem 5: repetitive planning work eats the planning cycle
Entering data, spreading annual numbers across months and copying assumptions between versions are necessary, but none of it is analysis. In a busy budget season it can crowd out the thinking that should matter most.
How JedoxAI solves it: the Planning Agent handles these tasks through simple instructions in plain English, such as distributing a figure or creating a forecast, and every change can be undone. Your team spends less time keying in numbers and more time challenging them. As we argued in the number one benefit of AI in EPM and FP&A, the real prize is time.
Problem 6: only one person understands the model
Every finance team has one: the person who built the model, knows how the data loads work and gets called when anything breaks. When they are on holiday, or leave, the business is exposed. It is one of the reasons Excel is the silent assassin inside modern finance.
How JedoxAI solves it: the Knowledge Agent gives step-by-step answers on how to use and configure Jedox, explains errors and can generate and review scripts in the Jedox Integrator. The Modelling Agent goes further, helping to build data integrations, models and reports on request: you can describe what you need and it proposes a structure for you to review. Knowledge is no longer locked in one person’s head, and changes no longer have to wait for IT.
Problem 7: systems, data and teams that don’t connect
Sales has one plan, operations another and HR its own headcount numbers. Data sits in the ERP, the CRM and a dozen spreadsheets. AI can’t give you a reliable answer if it can only see part of the picture.
How JedoxAI solves it: Jedox brings your systems, data sources and teams together in one model, which is exactly what makes its AI useful. JedoxAI works across finance and operational data, so the answer to “why is margin down?” can draw on sales volumes, pricing and costs together. Read more about integrated business planning.
Problem 8: “is it safe to put our numbers into AI?”
This is often the first question a CFO asks, and rightly so. Pasting financial data into a public AI tool is a real risk.
How JedoxAI solves it: JedoxAI runs on Microsoft Azure OpenAI Services with a dedicated, customer-specific model. According to Jedox, your data is isolated and encrypted, it is not shared with OpenAI or other third parties, and your prompts and data are not used to train external models. The roles and permissions you already have in Jedox still apply. If you want to use your own AI assistant, such as Claude or Copilot, an MCP connects it to Jedox under the same controls. Our guide what is an MCP? explains how.
The problem AI can’t solve on its own
AI is only as good as the data and model underneath it. If your structures are inconsistent or your data is poor, AI will give you fast answers you can’t trust. Stephen Hambling calls this playing AI Jenga with your FP&A. Getting the foundations right comes first, and our guide to fixing bad data in 5 steps is a good place to start.
How Kybos helps
As a Jedox Platinum Partner with accountancy-qualified consultants, Kybos builds the clean, well-structured models that make JedoxAI work. We cover implementation, training and support. If you already use Jedox, a Jedox health check will show how ready your model is for AI. You can also watch our Jedox agentic AI demo.
See JedoxAI working with your numbers
The quickest way to judge any AI is to see it answer your own questions. Arrange a 30-minute Jedox demo, or book a call with a Kybos finance consultant to talk through where AI could save your team the most time.
Frequently asked questions
What is JedoxAI?
JedoxAI is the AI built into the Jedox planning platform. It combines machine learning forecasting, natural-language questions and AI agents to help finance teams plan, analyse and report faster.
What problems does JedoxAI solve?
JedoxAI cuts the time it takes to get answers from your data, explains why numbers have changed, produces explainable forecasts, flags anomalies early, automates repetitive planning tasks and reduces reliance on a few model experts.
What are the JedoxAI agents?
JedoxAI includes a Reporting Agent for questions and insights, a Planning Agent for planning tasks, a Knowledge Agent for guidance and scripting help, and a Modelling Agent for building integrations, models and reports. Availability of each agent depends on your Jedox version.
Can I ask JedoxAI questions in plain English?
Yes. You can type questions such as “show me only actuals for 2022 to 2025” and JedoxAI updates the view for you, with no formulas or pivot tables needed.
Is my financial data safe with JedoxAI?
JedoxAI runs on Microsoft Azure OpenAI Services with a dedicated model for each customer. Jedox states that data is isolated and encrypted, is not shared with OpenAI or third parties, and is not used to train external models. Your existing Jedox permissions still apply.
Does JedoxAI replace the finance team?
No. JedoxAI does the first pass: gathering data, producing a baseline forecast and drafting the explanation. Your team still applies the judgement, challenges the numbers and advises the business. It gives them more time to do that well.
Do I need clean data to use JedoxAI?
Good data makes a big difference. AI is only as reliable as the model and data underneath it, so a well-structured Jedox model is the best starting point. A Kybos Jedox health check can tell you how ready you are.
Can I use Claude or Copilot with Jedox?
Yes. Using an MCP (Model Context Protocol), AI assistants such as Claude or Microsoft Copilot can securely read and work with Jedox data under your existing permissions.
How can I see JedoxAI in action?
Arrange a live Jedox demo with Kybos, or watch our Jedox agentic AI demo video on the Kybos website.
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.


