In recent years, Italy has digitalised much of its financial cycle. In 2025, the Italian Revenue Agency processed 2.4 billion electronic invoices (FiscoOggi, 2026). Electronic payments reached €518 billion, accounting for 45% of consumer spending, while bank transfers and direct debits rose to 17% (Innovative Payments Observatory, Politecnico di Milano, 2026). Recurring payments have become widespread across utilities, insurance, consumer credit, Buy Now, Pay Later (BNPL) services, and subscriptions. However, one process has lagged behind: credit management when payments are not received.
Late payments are a structural issue. According to CRIBIS, as of June 2026, only 42% of Italian companies paid their invoices on time (CRIBIS via Teleborsa, 2026).
For banks, fintechs, utilities and insurance companies, however, the challenge goes beyond individual invoices. It involves portfolios containing thousands or even millions of accounts, each with its own history, risk profile and customer relationship value.
Yet, in many organisations, credit management still relies on static rules, standardised payment reminders, infrequently updated customer segmentation and the transfer of cases to external service providers with incomplete information. The stage at which the amount of value recovered is determined is often the least digitalised part of the entire financial cycle.
In recurring payment business models, the debtor is not simply a counterparty to disengage from, but a customer to retain.
Some customers have simply forgotten a payment deadline, others are experiencing temporary financial difficulties, some dispute the amount due, while others have no intention of paying.
Treating all these situations in the same way creates a double cost: unrecovered debt and lost customers.
When discussing AI in debt collection, the first thing that often comes to mind is a chatbot sending payment reminders. However, this is only part of the picture.
Generative agentic AI can support credit management teams throughout the entire process, rather than simply managing customer communications.
In practice, AI agents can:
Most importantly, they learn. Every interaction generates data that can improve subsequent strategies. This is the difference between automating an individual task and automating a process capable of continuously optimising itself.
For banks, utilities and fintech companies managing large volumes of accounts, agentic AI fundamentally changes the economics of credit management.
Until now, personalised recovery strategies have generally been economically viable only for high-value accounts, while the rest of the portfolio has been managed through standardised processes.
With AI agents, every account can receive a tailored approach, regardless of scale.
This creates three concrete benefits:
Credit management professionals are not replaced. Instead, their role shifts from operational execution to supervision, rule-setting and the management of more complex or sensitive cases.
This is the approach behind JustSolve: an AI agent platform for credit management, integrated in real time with customer data and monitored through dashboards that provide visibility into every analysis, action and recovery outcome. The platform is designed to scale to millions of interactions per month.
Unlike other applications of AI in finance, AI in credit management does not operate exclusively in the back office: it communicates directly with people, sometimes when they are experiencing financial difficulties.
This makes certain principles non-negotiable.
The creditor retains control over the tone of communication. The rules governing what AI agents are authorised to propose, including repayment extensions, amounts and deadlines, must be explicit and verifiable.
Every interaction must be traceable, and more sensitive cases must be escalated to a human operator.
These safeguards must be accompanied by transparency towards debtors, in accordance with the relevant obligations under the EU AI Act applicable from 2 August 2026, as well as strict compliance with data protection legislation.
For large organisations, governance and consumer protection are not obstacles to innovation: they are essential conditions for its adoption.
Credit management is still often perceived as a cost centre to be outsourced or reduced.
With agentic AI, it can become a strategic driver of liquidity, customer intelligence and relationship management.
For those responsible for credit management in banks, fintech companies or utilities, the starting point is to consider three questions:
JustSolve is a Milan-based startup that has developed an AI agent platform for credit management, helping businesses recover more while spending less.
JustSolve’s AI agents analyse and segment credit portfolios, define tailored strategies for individual accounts and manage the entire debt recovery process across multiple communication channels, from negotiating repayment plans to issuing formal payment demands. The platform can typically be integrated into existing systems in less than a week.
Fonti
Article by:
Alessandra Calvi, Business Associate, JustSolve