Stop your chatbot from promising refunds it can't deliver.
Policy enforcement middleware for customer support AI. Sit between your LLM and your customer. Audit-ready. $500-5K/month.
Every LLM-powered CX deployment is one wrong answer from catastrophe.
The Moffatt v. Air Canada ruling made it case law: when your chatbot hallucinates a policy, your company is contractually bound. Decagon, Sierra, and Ada charge Fortune 500 customers $100K-$1M/year. Mid-market is wide open -- and more exposed.
Moffatt v. Air Canada
In 2024, a Canadian tribunal ruled Air Canada liable for a bereavement-fare policy its chatbot invented. The company argued the bot was a "separate legal entity." The tribunal disagreed. Your chatbot speaks for your company -- legally.
One answer, a million refunds
A single hallucinated promise, broadcast to every customer asking a similar question, becomes a contractual obligation you didn't agree to. CX AI hallucinations don't fail in isolation -- they fail at scale.
EU AI Act: high-risk
Customer-facing AI systems that make material decisions fall into the EU AI Act's high-risk classification, requiring risk management, audit logs, and human oversight. Fines reach 7% of global revenue. You need the paperwork before the regulator knocks.
Drop-in middleware. Every message checked before it ships.
The Guardrail sits between any LLM and your customer. Your model drafts a response. The Guardrail scores that response against your policy manifold. Anything outside the allowed region gets caught, redirected, or escalated -- with a full audit log attached.
Customer
Asks a question via chat, email, or voice
Your LLM
OpenAI, Anthropic, local, or any API
CX Guardrail
Policy check · refund ceiling · escalation · audit
Customer
Receives a safe, on-policy response
Policy match
Every proposed response is scored against your actual written policies -- not a vibes-based guardrail prompt, but a manifold check against your versioned policy files.
Refund ceiling
Hard caps on any monetary commitment the bot can make autonomously. Anything above the threshold requires human approval -- non-negotiable.
Escalation trigger
Legal language, distressed sentiment, regulated domains, or ambiguous requests are flagged and handed to a human agent before the bot answers.
Audit log
Every decision -- approved, rewritten, blocked, or escalated -- is cryptographically logged with the input, the draft, the policy hit, and the final output. Regulator-ready.
Three tiers. No surprises. Cancel anytime.
All tiers include the full audit log, policy-as-code authoring, and integration with any LLM. Scale up or down as your volume changes.
For small teams piloting LLM-powered support and wanting basic liability coverage.
- Up to 10K messages per month
- Standard policy templates (refunds, returns, shipping, eligibility)
- Email support, 1 business-day SLA
- Monthly audit log export (CSV + JSON)
- Works with any LLM API
- <100ms added latency
For scaling CX teams who need custom policies and real-time oversight.
- Up to 100K messages per month
- Custom policy authoring (YAML or JSON, versioned in your repo)
- Slack integration for escalations
- Real-time dashboard with block/rewrite/escalate counters
- Weekly audit reports delivered to your compliance inbox
- Everything in Starter
For regulated industries, multi-brand deployments, and teams needing a contracted SLA.
- Unlimited message volume
- White-glove onboarding (policy migration, LLM integration, QA)
- Dedicated Slack channel with our engineering team
- Custom integrations (Zendesk, Salesforce, Intercom, bespoke)
- Written SLA with uptime guarantees and response windows
- Everything in Growth
Outcomes, not features.
The Guardrail is boring on purpose. It sits in your stack and stops specific, measurable, expensive failures.
- Stop refund hallucinations before they reach the customer
- Satisfy EU AI Act high-risk CX documentation requirements
- Cryptographic audit trail for every decision your bot makes
- Integrate with any LLM in under a day (one API call, one header)
- Policy-as-code you version in git alongside the rest of your infra
- Human-in-the-loop escalation paths that actually trigger when they should
- Real-time alerting when your bot tries to go off-policy
- Regulator-ready reporting without a compliance team on speed dial
Built on a governance framework we've been hardening for years.
Powered by SCBE-AETHERMOORE
The CX Refund Guardrail is a productized slice of our patent-pending 14-layer governance pipeline (USPTO #63/961,403). The same geometric security model that prices out adversarial attacks via H(d,R) = R^(d^2) now enforces your customer-support policies at runtime. Your policies become a high-dimensional manifold; every LLM output gets scored against it; anything in the forbidden region is caught.
The core framework is open-source on GitHub. The Guardrail is what happens when you take that framework, pre-configure it for customer-support policy enforcement, and wrap it in an API.
The questions we get asked first.
Does this work with any LLM?
Yes. OpenAI, Anthropic, Google, local models (Llama, Mistral, Qwen), any HTTP API. The Guardrail is model-agnostic -- it evaluates the output, not the weights. If your model can return text, we can check it.
How fast is it?
Less than 100ms added latency per message, measured at the 95th percentile. The policy manifold check runs in parallel with a deterministic refund-ceiling check, and the audit log is written asynchronously.
Can we bring our own policies?
Yes. Policies are YAML or JSON files, versioned in your repo alongside the rest of your infrastructure. Change control, code review, and rollback are git operations you already know. No policy is ever edited through a hosted dashboard without leaving an audit trail.
What happens on a policy violation?
Configurable per rule: block the message entirely, escalate to a human agent, or rewrite the response using a safe template. You choose the default, and you can override per policy. Every action is logged.
Is there a free trial?
Yes -- 30 days, up to 1,000 messages, with the full Starter feature set. Email us to get provisioned; no credit card required to start.