Payhawk Unveils Four AI Agents, Including a Procurement Agent That Cuts Purchase Time by 60%

Technology & AI: Payhawk introduces AI agents for spend management

Payhawk has rolled out four AI agents in its Fall '25 Product Edition to make business spending faster, tighter, and easier to control. The focus: procurement and finance operations that need outcomes, not just another chat window.

Co-founders Hristo Borisov (CEO), Boyko Karadzhov (CTO), and Konstantin Dzhengozov (CFO) are positioning AI as an execution layer that sits on top of existing controls, approvals, and policies-so managers keep oversight while teams move quicker.

What's new

Payhawk's AI agents operate inside your existing roles and permissions, enforce policies automatically, and log every step for audit. Early pilots show month-end closing can be two to four days faster.

• Procurement Agent: Streamlines purchase requests via a chat-like interface, applies budgets and policies, and routes for approval. Payhawk claims up to 60% reduction in time-to-purchase.
• Payment Agent: Supports payment execution within configured limits and approval workflows.
• Travel Agent: Aligns bookings with company travel policies and budgets.
• Financial Controller Agent: Assists with reconciliations, compliance checks, and close tasks.

As Borisov puts it: "Enterprises don't need more chat, they need outcomes." Dzhengozov adds: "AI is a layer on top of all the controls and all the frameworks that already exist."

Why managers should care

These agents focus on measurable improvements that leaders can track across procurement and finance. If you're accountable for spend discipline and cycle time, the impact is straightforward.

• Faster month-end close: 2-4 days quicker in early results
• Time-to-purchase: up to 60% faster for standard requests
• Higher policy adherence: automatic checks reduce manual reviews
• Lower approval latency: routing and nudges reduce bottlenecks
• Better audit readiness: complete logs of actions and decisions

How it works for procurement

Employees request what they need in a chat-like flow. The Procurement Agent maps the request to the right cost center, budget, and policy, then advances it to the correct approver-no scavenger hunt for forms or templates.

The result is fewer back-and-forth messages, fewer exceptions, and more spend captured under policy without slowing the business down.

Governance, risk, and compliance

Managers get control without extra overhead. Agents operate within pre-set permissions, enforce policy limits, and keep a full activity log for audits. For teams aligning to recognized standards, this approach supports security and accountability practices found in frameworks like NIST AI RMF and ISO/IEC 27001.

Implementation playbook (30-90 days)

• Week 1-2: Map top 5 spend scenarios (low-risk purchases, travel bookings, recurring vendors). Define approval limits and exception rules.
• Week 3-4: Pilot the Procurement and Financial Controller Agents with a single business unit. Track cycle time, exception rate, and approval latency.
• Week 5-8: Expand to Payments and Travel. Integrate with ERP and HRIS for roles, cost centers, and budgets.
• Week 9-12: Scale company-wide. Lock in dashboards and alerts. Refresh policies to match what the agents can automate.

KPIs to own

• Days to close
• Time-to-purchase (request to PO)
• Policy compliance rate and exception volume
• Maverick spend as a percentage of total spend
• Manual journal entries per close
• Approval latency (by department and approver tier)

Practical guardrails

• Least-privilege access for agents; review role mappings quarterly
• Human-in-the-loop for high-value or high-risk transactions
• Clear exception handling and escalation paths
• Audit trails on by default; immutable logs retained per policy
• Vendor due diligence tied to data residency and security certifications

Industry momentum

The rise of agentic AI in procurement will be a key theme at Procurement & Supply Chain LIVE London 2025, including workshops featuring leading providers such as GEP. Leaders in fintech will also explore practical AI outcomes at FinTech LIVE London.

Bottom line for leadership

If your goal is speed with control, AI agents that act inside your governance model are worth a pilot. Start with standard purchases and month-end workflows, measure the lift, then scale what works.

For managers building skills and playbooks for AI-driven operations, explore curated resources for finance and procurement teams here: AI tools for finance.