Pricing & proof of value
Predictable tiers licensed by the number of LLM models using the product. Use the worksheet below for an editable planning scenario, then contact us to scope an assessment of your models and data flows.
Team
For platform teams running a small number of LLM-backed apps.
- Up to 2 LLMs protected
- Managed SaaS deployment
- Three upstream providers
- Standard policy pack
- OpenAI / Anthropic compatibility
- Email support (30% of license cost)
Business
For organizations rolling out AI across multiple business units.
- Up to 10 LLMs protected
- Single-tenant deployment option
- All upstream providers
- Custom policies, shadow mode
- SIEM / SOAR export
- SSO + SCIM, role-based access
- Priority support SLA (30% of license cost)
Enterprise
For regulated environments and the highest-volume deployments.
- Unlimited LLMs protected
- Air-gapped on-prem deployment
- Continuous red-teaming included
- Custom SIEM log formats
- Dedicated solutions engineer
- Quarterly assurance report
- Custom classifier fine-tuning
Risk-reduction ROI
Explore an illustrative financial scenario for inserting AI.Guard in front of your existing LLM traffic. It is not a measured outcome; an agreed evaluation can establish results using your own data.
Illustrative risk inputs
Edit these planning assumptions to explore a scenario; they are not MOAI-measured data.
Prompt-injection events, data leaks, or policy violations requiring formal response.
A user-editable planning assumption only; it is not a pilot result, benchmark, or guarantee.
Illustrative annual scenario
Baseline annual exposure
$540,000
Modeled residual exposure
$81,000
Modeled gross savings
$459,000
* Illustrative scenario only. Inputs and outputs are assumptions for planning and do not report measured AI.Guard performance. Any evaluation results would be scoped and measured using your traffic and agreed criteria.
A structured evaluation, if agreed
We don't do blind trials. If an evaluation is a fit, scope, access, timing, and success criteria are confirmed with your team before it begins.
Scope & deploy
Agree the model, request path, deployment boundary, data classes, and access needed for the evaluation.
Tune & enforce
Configure a limited set of confidential data types for masking and test agreed scenarios against representative traffic.
BVA report
Review observed detections, policy decisions, residual risks, and rollout questions against the agreed criteria.
Success criteria are agreed up front
- • p99 latency measured against an agreed baseline
- • Quantified detection rate against your chosen attack catalog
- • Documented reduction in residual risk for two named scenarios
- • Clean rollback path if you choose not to continue
Email request only; a meeting is not confirmed until MOAI Labs replies.