About
Alice (formerly ActiveFence) is a comprehensive security, safety, and trust platform purpose-built for the GenAI era. Trusted by more than 50% of the world's largest online platforms — covering 3B+ users across 120+ languages — Alice delivers a layered defense system through its flagship WonderSuite product family. WonderBuild enables pre-launch stress testing, helping teams identify vulnerabilities in AI models and agents before deployment. WonderFence provides dynamic runtime guardrails that enforce policy alignment, detect prompt injections, and block jailbreaks in real time. WonderCheck delivers ongoing automated red-teaming to surface model drift, emerging risks, and compliance gaps in production environments. Powered by a decade of real-world adversarial intelligence research and billions of continuously updated toxic examples, Alice's data moat enables classifiers that outperform industry benchmarks. Its adaptive policy alignment engine can be tuned to match specific regulatory requirements and risk tolerances across every content modality. Alice serves industries including child-facing products, financial services, healthcare, insurance, and legal services. It is used by leading foundation model labs — including Amazon AGI and Cohere — to accelerate model release cycles while maintaining responsible AI governance. The platform addresses critical challenges like unintended GenAI behavior, application-level exploitation, governance gaps, and reputational risk from AI failures.
Key Features
- WonderBuild – Pre-Launch Stress Testing: Simulates adversarial attacks on AI models, apps, and agents before deployment to identify vulnerabilities and ensure responsible, resilient release.
- WonderFence – Dynamic Runtime Guardrails: Monitors live AI applications in real time, blocking prompt injections, jailbreaks, and policy violations to keep interactions safe and brand-aligned.
- WonderCheck – Continuous Automated Red-Teaming: Runs ongoing evaluations in production to detect model drift, surface emerging threats, and prioritize risk remediation without manual intervention.
- Adversarial Intelligence Data Moat: Backed by a decade of real-world threat research and billions of toxic examples, Alice's proprietary dataset powers classifiers that outperform industry benchmarks.
- Adaptive Policy Alignment: Allows organizations to customize safety coverage across modalities, regulatory regimes, and risk tolerances to keep every interaction compliant and trusted.
Use Cases
- Testing foundation models and AI agents for safety vulnerabilities before public release to ensure responsible deployment.
- Enforcing real-time content policies on AI-powered chat applications to prevent jailbreaks, prompt injections, and harmful outputs.
- Continuously monitoring production LLM deployments for model drift and emerging risk patterns without disrupting live services.
- Meeting regulatory compliance requirements in healthcare, financial services, and legal sectors by aligning AI outputs with jurisdiction-specific content policies.
- Protecting user-generated content platforms from harmful, abusive, or illegal content at scale across 120+ languages.
Pros
- Battle-Tested at Massive Scale: Protects 3B+ users across 120+ languages and 1B+ daily AI-human interactions, proving reliability at enterprise and internet scale.
- Full AI Lifecycle Coverage: Addresses safety at every stage — pre-launch, runtime, and post-deployment — eliminating blind spots that point-solution tools miss.
- Trusted by Leading Foundation Model Labs: Used by Amazon AGI, Cohere, and other top AI organizations, validating its depth of safety expertise and quality of adversarial coverage.
- Regulatory and Compliance Ready: Supports industry-specific requirements across healthcare, finance, legal, and child-safety sectors with tunable policy enforcement.
Cons
- Enterprise-Only Pricing: No self-serve or free tier is available, making it inaccessible for smaller teams, startups, or individual developers exploring AI safety tooling.
- Requires Sales Engagement: Prospective customers must request a demo to get started, adding friction compared to platforms with instant sign-up and trial access.
- Complexity for Smaller Deployments: The breadth of WonderSuite's capabilities may be more than needed for simple AI use cases, adding implementation overhead for smaller-scale projects.
Frequently Asked Questions
What is Alice (formerly ActiveFence)?
Alice is an enterprise AI security and safety platform that provides pre-launch stress testing, real-time runtime guardrails, and continuous automated red-teaming to protect GenAI apps, agents, and UGC platforms from adversarial threats.
What is WonderSuite?
WonderSuite is Alice's flagship product family consisting of three components: WonderBuild (pre-launch testing), WonderFence (runtime guardrails), and WonderCheck (ongoing production red-teaming).
Who uses Alice?
Alice is used by large enterprises, foundation model labs (including Amazon AGI and Cohere), and platforms in regulated industries such as healthcare, financial services, insurance, legal services, and child-facing products.
What threats does Alice protect against?
Alice protects against prompt injections, jailbreaks, model exploits, unintended GenAI behavior, data leakage, governance gaps, and reputational risks from harmful or non-compliant AI-generated content.
How does Alice's data advantage work?
Alice has accumulated a decade of real-world adversarial intelligence research and billions of continuously updated toxic examples. This proprietary dataset enables more accurate classifiers and stronger red-teaming capabilities than standard benchmarks.
