
Banking & payments
Assisted credit assessment · Disputes, chargebacks and complaints · Reconciliations and exceptions
- Data Foundation
- Data Intelligence
- Agents
- Workspace

A credit application, an insurance claim and a payment dispute all run through documents, data and people who decide, but each has its own rules and its own answer to who is accountable for a mistake. That is why we work alongside people who know the trade.
Under different names, many use cases call for the same work on documents and data, which is why what we build for one process also serves the next.
| Type of work | Where we find it | Modules most involved |
|---|---|---|
| Cases assessed on documents | Where we find itAssisted credit assessment, in banking · Claims from notification to settlement, in insurance · Admissions and documentation, in healthcare | Modules most involvedAgents, Company Brain, Data Foundation, Workspace |
| Complaints, disputes and refunds | Where we find itDisputes, chargebacks and complaints, in payments · Billing complaints and corrections, in utilities · Orders, returns and refunds, in retail | Modules most involvedAgents, Workspace |
| Reconciliations and controls | Where we find itReconciliations and exceptions, in banking · Accounting close and reconciliations, in insurance · Completeness and coding of services, in healthcare | Modules most involvedData Foundation, Data Intelligence, Agents |
| Company knowledge and support | Where we find itProduct knowledge for AI agents, in insurance · Assistant for field engineers, in utilities · Company knowledge assistant, in every industry | Modules most involvedCompany Brain, Workspace |
| Forecasting | Where we find itDemand and stock planning, in manufacturing · Store-level demand forecasting, in retail · Demand and consumption forecasting, in utilities | Modules most involvedData Foundation, Data Intelligence |
Many internal questions, in a bank's credit department as much as in a manufacturer's technical support team, are answered by documents that already exist, such as a procedure or the latest version of a contract. The assistant answers from those documents and gives the source of every answer. We build it on the Company Brain, which holds what the company knows.
An example: in a mid-sized company, the HR department gets the same questions every week about holiday, leave and expense claims. The head of HR chooses which policies count as sources. The assistant cites them in its answers and hands the questions that call for judgement, such as an exception to the rule, back to the head of HR.
Permissions stay as they are in the source systems, so everyone sees only what they could already open. When the documents concern employees, GDPR and employment law apply too, starting with Article 4 of the Italian Workers' Statute (Statuto dei lavoratori) on remote monitoring. If internal documents are out of date or contradict one another, the assistant repeats their mistakes, so the first job is to decide which sources count.
The use cases on these pages are possible projects, which we assess with the process owner before choosing one. We usually start with a Mapping, which chooses where to act, or with Deployment if the process has already been decided. The second process then reuses the data and rules of the first. A Group Partner who knows your industry joins the first meeting.
If you already have one in mind, we will discuss it with someone who knows your industry; if not, the first meeting is there to find it.