Creditinfo Estonia and Itronauts have jointly developed a service that lets Creditinfo customers ask questions about companies in ChatGPT, Codex and Claude, using everyday language. The project went from kickoff to public beta in eight weeks, and development continues with additional features.
For Creditinfo, this opens up another way to serve its customers. They can now work with Creditinfo’s business data in the same AI tools they already use to analyse information and prepare decisions.
A new way to access Creditinfo’s data
The solution is built around an MCP server. MCP, short for Model Context Protocol, is a standard that allows AI tools to request information from external systems. In this project, the MCP connection makes it possible to retrieve information directly from Creditinfo through ChatGPT, Codex and Claude.
Customers access the service through their Creditinfo accounts, with data availability determined by their permissions and subscription plan. This makes the new service a natural extension of their existing relationship with Creditinfo.
The solution supports more than 20 types of data query, covering company registration details, credit ratings, payment defaults, tax information and annual reports, among other areas. Dedicated workflows support common tasks: credit checks, company comparisons and explanations of risk signals.
What can customers do with it?
Users describe what they need in everyday language. The relevant information is then retrieved from Creditinfo and used to prepare an overview.
Consider a sales manager whose new customer wants 30-day payment terms on an €8,000 order. Before agreeing, the manager needs to check the company’s background and review several indicators.
They can ask:
“Run a background check on Example Ltd. Include its credit rating, payment defaults and tax arrears, and compare an €8,000 order with Creditinfo’s recommended credit limit.”
The resulting overview highlights relevant risk signals and identifies the sources and dates of the data. The sales manager can use it to inform their decision and ask follow-up questions in the same conversation.
The same approach can help customers compare the financial position of potential suppliers or investigate changes in an existing business partner’s data.
The decision stays with the person. The service reduces the work involved in finding and bringing together the information they need.
How the data is presented matters
An important part of the development work was defining how information should be interpreted and presented.
“No payment defaults found” needs to be clearly distinguished from “The payment default check could not be completed.” If several companies have similar names, the system needs to establish which one the user means before proceeding.
Rules for presenting data, identifying sources and handling unsuccessful queries therefore received particular attention. These details directly affect how useful the overview is to someone deciding what to do next.
Where else could MCP be useful?
The Creditinfo project is one example of making a company’s data available through AI tools. The same approach could also work in other industries.
For example, a customer of a B2B IT equipment store could ask in ChatGPT:
“We need laptops and compatible docking stations for ten new employees. Our budget is €1,500 per workstation, and everything needs to arrive by next Friday. What are our options?”
If the store makes product specifications, stock availability, customer-specific pricing and delivery estimates available through MCP, the tool could use that information to compare suitable configurations. Buyers could start by describing their requirements, while the store gains another channel through which to serve them.
For an internal use case, a sales team leader might ask: “Which customers have we sent quotes to in the past month without following up?”
An AI tool connected to the company’s sales system could review quotes and recorded communication history, then compile a list of customers to contact.
Both examples depend on access to the information needed for the task, within the user’s permissions. MCP provides a way to establish that connection.
The public beta is live, and development continues
Through the combined work of Creditinfo and Itronauts, the service reached public beta in eight weeks. The teams are now continuing to develop additional features.
You can explore the service and find setup instructions on the Creditinfo AI connection website.
Could your customers or employees benefit from accessing your company’s information directly through their AI tools? Itronauts develops AI solutions and MCP servers that connect business data and workflows to those tools. Get in touch — we can help you identify a useful starting point and build it.