Launch SKU: Agent Governance for Regulated Workflows

Ideal Customer Profile (ICP)

Primary ICP: Mid-market and enterprise product teams deploying LLM agents into regulated workflows where decisions must be explainable, controllable, and exportable for compliance.

Target segments

• Fintech, insurance, healthcare, and public-sector vendors operating under audit obligations.
• B2B SaaS platforms embedding agentic automation into customer-facing operations.
• Platform/security teams responsible for production guardrails across multiple agents and tools.

Buying triggers

• Need to prevent unsafe tool calls before execution.
• Need to block or quarantine sensitive data egress in real time.
• Need machine-readable audit logs for internal controls and external regulators.

Top 3 Launch Use Cases

1. Tool-call control
   • Enforce allow/deny policy on agent tool invocations (by tool, argument class, identity, and context).
   • Support deterministic authorize/deny/quarantine outcomes before execution.
2. Data egress guard
   • Inspect outbound payloads and tool arguments for regulated/sensitive content classes.
   • Quarantine or redact risky outputs to reduce exfiltration and policy breaches.
3. Audit export
   • Export decision and policy events to SIEM/data-lake destinations.
   • Preserve evidence chain for compliance reviews and incident postmortems.

Pricing Unit (Launch Recommendation)

Use a hybrid model to align value with both usage and organizational footprint:

• Primary meter: Per 1,000 decisions (authorize/deny/quarantine evaluations).
• Platform floor: Per tenant base fee (control plane, policy lifecycle, retention).
• Optional packaging: Per agent bundles for customers that prefer predictable budgeting.

Why this works

• Per-1k decisions tracks variable risk-screening load and scales with adoption.
• Per-tenant floor captures fixed compliance and governance value.
• Per-agent option simplifies procurement for teams with stable agent counts.

Success SLOs (Launch)

1. p95 authorize latency: ≤ 120 ms at launch, with target ≤ 80 ms by end of quarter.
2. False quarantine rate: ≤ 1.5% for approved policy packs on reference traffic.
3. Audit export SLA: 99.9% successful delivery within 5 minutes of decision event creation.

12-Week Roadmap Tied to Revenue

Timeline	Deliverable	Revenue Link	Monetization Impact
Week 1–4	API + gateway	Fastest path to production enforcement in customer pilots.	Unlock paid design partners billed per 1k decisions; start usage revenue immediately once traffic is routed.
Week 5–8	Policy lifecycle + audit export	Converts pilots to compliance-ready deployments.	Enables per-tenant platform fee and expansion from security buyer to compliance budget owner.
Week 9–12	SDK + 2 reference integrations (OpenAI agent + browser agent)	Reduces integration friction and broadens addressable agent stack.	Drives seat/agent expansion inside existing tenants and shortens sales cycles for new logos.

Commercial Packaging (Suggested)

• Launch (Design Partner): discounted per-1k decision rate + limited per-tenant fee, with implementation support.
• Growth: standard per-tenant platform fee + tiered per-1k decision pricing.
• Enterprise: volume-committed per-1k pricing, advanced retention/export controls, and custom SLA addenda.

Definition of Launch Readiness

Launch SKU is ready when:

• API + gateway are GA for enforcement in at least two production-like customer environments.
• Policy lifecycle supports versioning, staged rollout, and rollback.
• Audit exports meet 99.9% / 5-minute SLA for two consecutive weeks.
• Reference integrations demonstrate end-to-end value on both OpenAI and browser-agent workflows.