Forrester: A Third of B2B Payments Will Run on AI Agents by Year End

Forrester projects that roughly one-third of B2B payment operations will rely on AI agents by the end of 2026, automating invoice matching, payment reconciliation, cross-border FX optimization, and dispute resolution — tasks that have historically consumed significant finance-team headcount.

The Productivity Case, With Real Numbers

McKinsey research finds businesses using AI agents for multi-party reconciliation and real-time liquidity management see a 20-40% reduction in processing time and a 30-50% decrease in backlog — a meaningfully larger gain than most enterprise automation initiatives report, likely because reconciliation is a genuinely rules-heavy, pattern-matching task well suited to agentic AI's current capability level.

Where This Fits in the Broader Agentic Payments Shift

• AI is moving into risk scoring, fraud checks, and support as core payment infrastructure capability, not a peripheral add-on feature
• Firms are increasingly building multi-agent systems where several agents or AI services coordinate across a workflow, rather than deploying single-purpose bots
• The IMF has separately noted agentic AI could improve financial inclusion by reducing search costs and information asymmetries, letting users compare financial products and terms more effectively

For finance leaders, the practical question this raises isn't whether to pilot AI agents in reconciliation — the productivity case is now well-documented — but how to govern agent decision authority in a function where errors carry direct financial and audit consequences.