Why Collections Still Needs Human Judgement
Crest is developed and supported by Channel Digital Technologies, part of the wider Channel group. Channel has been operating for over a decade in receivables finance and credit management, and is FCA regulated, ISO and Cyber Essentials accredited. Crest itself was originally built to support large, multi-country operations managing complex invoice flows and funding structures. Today, it has evolved into a leading enterprise AR automation solution and accounts receivable automated platform, designed to deliver end-to-end AR automation for modern finance teams.
Purpose-built as AR automation for global businesses, Crest is a robust, production-grade platform processing millions of documents annually while supporting complex, cross-border financial operations.
The team behind it combines deep domain knowledge in receivables finance with hands-on technical expertise, and the platform is hosted on AWS with enterprise-grade security and resilience built in from the ground up.
In 2026 the language around Accounts Receivable is dominated by automation. Vendors talk about “autonomous finance”, predictive payment models and collections engines that promise to remove almost, if not all, manual effort from the order-to-cash cycle.
For finance leaders the idea is understandably appealing. If software can prioritise accounts, send reminders and guide collectors on what to chase next, AR should become far more efficient. And in some areas this undoubtedly makes sense, with significant upside and very little downside.
Routine collections activity is easy to automate. Sending reminders, escalating overdue balances, predicting payment dates and organising collector workloads are repetitive tasks where systems perform well. The challenge appears when real situations inevitably arise.
In large organisations, especially those operating across multiple countries, overdue invoices are rarely just about payment behaviour. They often involve disputes, operational issues or commercial relationships that software cannot easily interpret, at least not in 2026.
- A payment might be delayed because a timesheet was approved late.
- Procurement may require additional documentation before releasing funds.
- A long-standing client might be temporarily slow but still strategically important.
In these situations the correct action often depends on judgement and relationships built over time. Experienced collectors bring context that systems struggle to understand. They know which customers respond to pressure and which require patience. They can judge whether a dispute is genuine, whether a reminder will help, or whether escalation risks damaging a valuable relationship.
These decisions rarely follow clean rules. What good AR systems should do is allow teams to apply different approaches to different customers. A global enterprise client may require a careful, relationship-led collections approach, while smaller accounts might follow a far more automated reminder cycle.
Automation works well when the task is repetitive and predictable. Collections often aren’t. The strongest AR teams therefore use automation to remove administrative effort such as reminders, prioritisation and communication tracking. That frees collectors to focus on the accounts where thinking, experience and commercial judgement actually matter.
Technology can handle the routine work. But collections will still need people who understand the situation behind the invoice.
