FOR CX TEAMS
Deploy AI Agents
Without Losing Control of the Customer.
Clyro enforces your business rules at runtime, so a support agent will not invent a refund policy, approve an order outside your limits or escalate a problem only after the customer complains.
The problem CX teams keep running into
Klarna deployed AI agents that handled 2.3 million customer conversations in a single month, roughly the workload of 700 full-time support agents. Within a year the company was hiring support staff back. The volume was never the issue. The agents could not handle edge cases, upset customers or knowing when they were out of their depth.
The model was not the weak point. With no rules around it, a support agent will invent a refund policy or approve an order no human would, and you find out when the customer complains.
What Clyro enforces, by default
Maximum refund amount: $500
The agent cannot override this.
Approved topics: returns, shipping, product info
The agent stays in its lane.
Escalation triggers: legal questions, complaints, edge cases
The agent hands off to a human.
Quantity limits: no single order above 50 units
No repeat of the 260 McNuggets order.
One YAML file, four rules, every conversation
Drop a policy file alongside your agent. Clyro checks every action against it before the agent responds.
# cx-guardrails.yaml
rules:
- id: refund-limit
condition: { field: refund.amount, operator: max_value, value: 500 }
action: block
message: "Refunds over $500 require manager approval"
- id: topic-restriction
condition: { field: response.topic, operator: in_list, value: [returns, shipping, product_info] }
action: allow
- id: quantity-cap
condition: { field: order.item_quantity, operator: max_value, value: 50 }
action: block
message: "Quantities over 50 require confirmation"
- id: escalation-trigger
condition: { field: request.category, operator: in_list, value: [legal, complaint, edge_case] }
action: escalate
message: "Routing to human agent"Pre-built CX policy templates ship with Clyro. Customize them or use the defaults.
What you get on day one
- Volume handled within bounds. The agent stops at the refund cap, the quantity cap, the topic boundary.
- Human escalation that fires reliably. Legal, complaints and edge cases route to a human, every time.
- An Agent Reliability Index per agent. Track which agents are drifting before customers do.
- Audit trail for every decision. Compliance and support leads see what shipped and why.