Rasmus Areskoug

Dunning is the process of following up on overdue invoices to ensure payment is received. It is a structured workflow that includes reminders, resolving issues that block payment, and escalating when necessary, all while keeping customer relationships professional and consistent. In SaaS companies, dunning works a little differently. Recurring revenue doesn’t always mean cash arrives on time, because unpaid invoices are often stuck in subscription workflows, billing questions, or internal accounts payable processes. Early-stage dunning confirms invoices are received and sent to the right person, while later stages involve more structured follow-ups, escalations, and actions tailored to different customer types. Effective dunning is more than sending reminder emails. Automation handles both routine follow-ups and the more complex tasks that arise when payments are delayed, but AI agents take it further. They can manage incoming emails, resolve disputes, capture missing information, and update systems automatically, making the entire dunning process and collections faster for finance teams.

Working capital management is a critical lever for finance leaders seeking to improve cash flow, operational flexibility, and business resilience. While many organisations focus on a single balance sheet figure, leading finance teams track multiple metrics to identify precisely where cash is tied up, including receivables, payables, inventory, and capital efficiency. In 2026, AI agents are increasingly helping finance teams optimise working capital by reducing delays in collections, resolving disputes, and providing actionable insights without adding headcount. This guide explores the seven key working capital metrics, how to calculate them, practical examples, and how AI agents can improve outcomes—especially by reducing Days Sales Outstanding (DSO).

Bad debt, doubtful allowances, and impairments are more than accounting concepts; they are practical levers that allow finance leaders to anticipate risk, protect cash flow, and make informed, confident decisions. These concepts work together as an ongoing credit risk management framework. By understanding how these concepts relate, spotting early warning signals in collection behaviour, and implementing a structured decision-making framework, finance teams can move from reactive write-offs to proactive management. Bad debt is the amount confirmed as uncollectible after reasonable recovery efforts fail, resulting in a write-off and a hit to profitability. Doubtful allowances are a proactive estimate of receivables unlikely to be collected, calculated using ageing, historical defaults, and customer-specific factors, reducing net AR without reversing revenue. Impairments are recognised when a receivable’s recoverable value declines due to credit deterioration, disputes, or customer financial difficulty, often before formal default, aligned with IFRS 9 expected credit loss thinking. Allowances predict portfolio-level loss, impairments flag specific exposures, and bad debt confirms loss once recovery is no longer realistic. Bad debt builds quietly through unresolved billing queries, missing documentation, slow dispute resolution, and broken payment promises. Early warning signs hide in email threads, CRM notes, and support tickets, not financial systems.

AI is not just automation, it spans three distinct approaches: rules-based automation, AI copilots, and AI agents; each suited to different types of work. AI agents represent a shift from assistance to execution, taking actions across systems, completing end-to-end workflows under human supervision and actually doing the work. Finance has adopted AI more cautiously than functions such as engineering, support, and sales. Finance is now well-positioned for agentic adoption, particularly in high-volume, repetitive, and measurable work. The highest-ROI uses of AI agents today are in operational finance, where these conditions are most common. The first and clearest high-ROI use cases are accounts receivable (billing support and collections automation) and accounts payable (invoice capture, coding, matching, and duplicate detection) Real value comes from agentic-native platforms that can safely act within systems, not legacy tools with chatbots rebranded as “AI-powered.”

AR teams managing hundreds or thousands of invoices per month hit a ceiling that process changes alone cannot fix. The constraint is not invoicing or reminder cadence; it is the volume of customer communication required before payment can proceed. Requests for invoice copies, PO validation, dispute resolution, and payment confirmation drive most delays. When these messages sit unanswered in finance inboxes, invoices age even when customers intend to pay. Workflow automation cannot manage inbound AR communication: ERPs and traditional AR tools track balances and send reminders, but they cannot read customer replies, resolve questions, or capture promise-to-pay commitments hidden in email threads. Platforms like Paraglide operate directly inside the finance inbox, using AI agents to manage customer replies, retrieve documents, capture promise-to-pay dates, and escalate only when human judgment is required, removing the communication bottleneck that drives DSO.

Decreasing DSO is about more than chasing customers faster — it's about building smarter, data-driven, and automated AR workflows. With AI agents managing reminders, resolving issues, and escalating intelligently, finance teams can collect faster, improve cash flow, and maintain strong customer relationships.

Kundreskontra är en central process för alla företag, men den har länge varit tung och arbetsintensiv. Med AI och automatisering kan företag nu frigöra tid, minska DSO, förbättra AR turnover och få bättre kontroll över sina åldersanalyser. Resultatet blir snabbare inbetalningar, starkare kassaflöde och en ekonomiavdelning som kan fokusera på det som verkligen driver värde.

Free cash flow conversion is a vital measure of financial strength because it shows how much of your reported earnings actually makes it into the bank. The biggest drag on FCF conversion is usually working capital, and within that, slow-paying customers. By managing accounts receivable tightly and reducing DSO, companies can unlock significant amounts of cash and transform weak FCF conversion into strong performance.

The role of the credit controller is evolving. Today, much of the job involves repetitive email and phone conversations with customers. Tomorrow, AI agents will handle these at scale, allowing humans to focus on the judgment calls, the disputes, and the exceptions that truly require their expertise. Credit controllers will not be replaced. They will be elevated, moving up the value chain to manage AI-driven teams of agents and focus on more strategic, value-adding work.

In large AR teams, overdue payments are usually due to inbox overload, not customer refusal to pay. High volumes of routine queries, missing POs, disputes, remittances, and invoice requests create delays that increase DSO, weaken forecasting, and stretch teams beyond capacity. AI agents for accounts receivable remove this bottleneck by automating high-volume finance communication. Agentic AR solutions can read and respond to emails, capture promise-to-pay dates, resolve disputes, match payments, and sync updates to ERP systems, turning unstructured inbox traffic into structured cash-flow data.