Robotic Process Automation (RPA)

Robotic Process Automation (RPA) is a category of software that automates repetitive, rule-based digital tasks by mimicking the actions a person would take in a user interface or through application APIs. A configured 'bot' can log into systems, copy data between applications, fill forms, move files, and trigger downstream steps without changing the underlying software. RPA emerged as a way to bridge systems that lack direct integrations, automating the manual hand-offs that sit between them.

Traditional RPA follows explicit, deterministic scripts: it does exactly what it is told and breaks when a screen layout, field, or workflow changes. It excels at high-volume, stable, well-defined processes such as data entry, reconciliation, and report generation, but it does not reason about context or handle ambiguity. Maintenance overhead from brittle scripts is a common criticism of older RPA deployments.

RPA is often contrasted with AI agents. Where classic RPA replays fixed steps, AI-driven automation can interpret instructions, adapt to variation, and decide which action fits a situation. In an AI-workforce product, agents that connect to a customer's tools and propose concrete actions occupy a similar 'get work done across systems' niche as RPA, but rely on language-model reasoning and human approval rather than recorded macros.