Gradient Labs · FinOps Profile

Gradient Labs Finops

FinOps view of Gradient Labs spend. Gradient Labs is an enterprise, outcome-based service: cost is driven by the volume of conversations the AI agent ("Otto") resolves across channels rather than by metered per-token or per-API-call rates. Because pricing is sales-quoted and not publicly itemized, cost allocation should be built around conversation volume, resolution/hand-off rates, and per-channel usage tracked from the API and webhooks.

Gradient Labs Finops is the FinOps profile for Gradient Labs on the APIs.io network, aligned with the FinOps Foundation Framework.

It defines 5 billable meters, billed in USD, on a monthly cycle, and pricing category outcome-based.

The profile maps 8 FOCUS columns for cost-allocation reporting.

Tagged areas include AI, Customer Support, AI Agent, Financial Services, and FinOps.

Category: AI and Machine Learning Pricing: Outcome-Based Billing: Monthly FOCUS v1.3
AICustomer SupportAI AgentFinancial ServicesFinOpsCost ManagementFOCUS

Framework Alignment

Framework
FinOps Foundation Framework
Data Spec
FOCUS v1.3

Charge Categories

UsagePurchase

FOCUS Columns

BillingCurrency
USD
ChargeCategory
Usage
InvoiceIssuerName
Gradient Labs
PricingCategory
Outcome-Based
ProviderName
Gradient Labs
PublisherName
Gradient Labs
ServiceCategory
AI and Machine Learning
ServiceName
Gradient Labs AI Agent

Meters

conversations_started
Unit: conversations
Conversations created via POST /conversations, a leading indicator of volume.
conversations_resolved
Unit: conversations
Conversations the AI agent concluded (conversation.finished webhook); the primary outcome-based cost driver.
conversations_handed_off
Unit: conversations
Conversations escalated to a human (conversation.hand_off webhook); tracks deflection rate and residual human cost.
agent_messages
Unit: messages
Outbound AI-agent messages delivered (agent.message webhook).
actions_executed
Unit: invocations
Business tools/actions the agent invoked (action.execute webhook or POST /tools/{id}/execute), which may incur downstream system cost.

Sources