Basware has introduced a Governed Autonomy Framework enabling finance teams to expand AI authority within customer-defined governance controls.
The Finland-based invoice lifecycle management (ILM) provider has announced an operating model designed to let organisations increase the scope of AI activity in finance functions while maintaining defined policies, approval thresholds and audit trails. The framework addresses a tension that has emerged as AI moves from an advisory role into direct execution of financial processes such as invoice review and exception handling.
From assistance to execution
According to a survey of global finance leaders, 92% of chief financial officers said they are already integrating AI into financial decision-making across some or virtually all processes. Separate research from PwC found that only 12% of business leaders report AI has delivered both cost and revenue benefits, while 56% say they have seen no significant financial benefit to date. Basware frames this gap as a consequence of systems that either automate heavily and risk compliance exposure when decisions cannot be traced, or apply tight controls that limit the benefits of automation.
The Governed Autonomy Framework sets out three levels of AI authority, each configured by the customer rather than fixed by the vendor: AI as adviser, where the system recommends and a person decides; AI as collaborator, where the system acts within agreed thresholds subject to human review; and AI as operator, where the system runs autonomously within defined controls, with every action logged and auditable. Organisations can move between these levels on a process-by-process basis as confidence in outcomes builds, without reconfiguring the underlying system.
Governance embedded in the workflow
Basware positions compliance as a foundation for the framework rather than an additional layer. As an example, the company describes a three-way match exception involving a USD 480,000 invoice in which the unit price sits 4.3% above the contracted rate, within a tolerance range some organisations permit. The system applies the customer’s pre-configured policy, factors in the supplier’s 36-month payment history, and either resolves the exception automatically with a recorded decision trail or escalates it with a structured recommendation, depending on the threshold set by the finance team.
The framework draws on data Basware has accumulated over more than four decades of invoice processing for more than 6,500 enterprise customers, spanning over 2.5 billion invoices and supporting more than one billion AI-driven actions annually.
Regulatory context
The announcement follows a period of expanding e-invoicing and reporting mandates across Europe, including France’s e-invoicing reforms, Poland’s KSeF framework, Germany’s B2B e-invoicing requirement, and the EU’s VAT in the Digital Age (ViDA) initiative. Basware states its compliance coverage extends to more than 60 country mandates, maintained as requirements change. The company positions this regulatory coverage as embedded within the AI system’s decision-making rather than applied separately to AI-specific features, with accountability for outcomes remaining with the organisations that configure the thresholds.
