MUMBAI, India, Aug. 31 -- Intellectual Property India has published a patent application (202631101804 A) filed by Saumen Bandyopadhyay on August 24, 2026, for A Predictive System For Forecasting Invoice Payment Delays And Dynamically Prioritising Accounts Receivable Collection Activities.
Inventor includes Saumen Bandyopadhyay.
The application for the patent was published on August 28, 2026, under issue no. 35/2026.
Abstract: The invention provides an event-driven system for forecasting invoice payment delay and maintaining an accounts-receivable task queue. Source connectors receive invoice, payment and account-state events. A canonical processor resolves invoice keys, suppresses duplicates and qualifies event time, freshness and schema. An incremental store updates only affected invoice feature states. A forecasting engine produces a payment-delay probability or interval with uncertainty, model-version and expiry metadata. A validity evaluator substitutes deterministic queue treatment when inputs or operating regime are unsupported. A constrained queue engine allocates tasks under worker capacity, connector capacity, account restrictions and contact-frequency limits. An orchestration interface dispatches a task with an idempotent token. A feedback engine correlates acknowledgements and subsequent payment or dispute events with the token and withdraws, retains or re-ranks the task. The architecture reduces redundant ledger processing, duplicate connector actions and stale queue entries while retaining auditable fallback operation
Disclaimer: Curated by HT Syndication.
