Veris AI Announces Availability on Google Cloud Marketplace to Deliver Reliable, Production-Ready AI Agents
Veris AI, the simulation and runtime platform for training, testing, and safeguarding AI agents, announces the availability of its platform on Google Cloud Marketplace. The listing enables enterprises building AI agents on Google Cloud to deploy Veris AI with streamlined procurement, simplified billing, and seamless integration with their existing Google Cloud infrastructure.
The Challenge: Making AI Agents Reliable Enough and Ready for Production
As enterprises accelerate the adoption of AI agents for mission-critical workflows, from customer service and supply chain management to executive assistance and financial operations, the gap between prototype and production-ready remains a major barrier. Static test sets and prompt engineering cannot capture the dynamic, multi-turn, stateful complexity of real agentic workflows. Organizations need a way to rigorously train, test, and govern their agents before and after deployment.
Veris AI solves this by simulating the full "world" an AI agent operates in, including users, tools, APIs, and edge cases, so teams can ship reliable agents to production with confidence and safeguard them once they're live.
What Veris AI Brings to Google Cloud Customers
With Veris AI now on Google Cloud Marketplace, enterprises can access the platform's core capabilities:
Veris Simulation
A high-fidelity simulation sandbox environment that trains and evaluates AI agents before deployment. Veris’s sandboxes are air-gapped and pre-loaded with simulated enterprise tools and users.
Veris generates thousands of realistic scenarios, including adversarial cases, tool failures, and complex multi-step scenarios, to surface problems before customers do. The platform simulates the entire environment an agent interacts with: users with realistic personas and behaviors, enterprise tools (Email, Slack, Jira, databases, CRMs, and more), APIs, and system states, all maintaining full statefulness throughout each episode.
Key functionalities:
- Experimentation and A/B testing: compare different models, tools, and architectures using a set of robust testing scenarios in Veris sandbox
- Safety and alignment testing: detect PII leaks, hallucinations, prompt injection vulnerabilities, and policy violations before going into production
- CI/CD integration: trigger simulation runs on every pull request to prevent regressions before they reach production
- Reinforcement Learning as a service: structured traces compatible with RLHF, RLAIF, and Reinforcement Fine-Tuning pipelines, with built-in prompt optimization
Key features:
- Interactive simulation environments: test AI agents in an air-gaped simulated environments that can handle multi-tool, multi-turn, conversations
- Probabilistic user simulation: simulated users with unique goals, personality traits, and frustration thresholds across text, voice, email and other modalities
- Stateful enterprise tool mocking: persistent simulations of tools like ERP, Slack, Jira, Postgres, CRMs, etc. where actions carry forward across turns
- Root cause analysis reports: comprehensive root cause analysis report with actionable recommendations on how to improve the agent using better prompts, harnesses, tool descriptions, and model fine-tuning
Veris Runtime
A production control plane that governs what an AI agent is allowed to do in real-time based on demonstrated competence. Unlike traditional guardrails that rely on static block/allow rules, Veris Runtime uses dynamic, competence-based routing and continuously improves the agent from real-world failures.
Key capabilities include:
- Competence Gate: routes incoming tasks based on the agent's proven ability, escalating unproven scenarios to humans
- Action Rollback: a universal undo for agent actions, combining internal state rollback with staged external calls to prevent irreversible mistakes
- Tool Firewall: evaluates each tool call against competence data and safety policies before execution
- Continuous improvement loop: automatically converts production failures into simulation scenarios for retraining
Together, Simulation and Runtime form a closed loop: simulate, deploy, monitor, improve, and re-deploy agents.
“Bringing Veris AI to Google Cloud Marketplace will help customers quickly deploy, manage, and grow the AI simulation sandbox platform for AI agents on Google Cloud's trusted, global infrastructure," said Dai Vu, Managing Director, Marketplace & ISV GTM Programs at Google Cloud. “Veris AI can now securely scale and support customers on their digital transformation journeys.”
Built for Enterprise Security and Flexibility
Veris AI is SOC2 Type II certified and supports flexible deployment options including Google Cloud VPC, on-premises infrastructure, or Veris AI Cloud. The platform is framework-agnostic and model-agnostic, working with any agent regardless of the underlying LLM, whether Gemini, GPT, Llama, Mistral, or others as well as any framework including Google ADK, LangGraph, OpenAI SDK, or custom implementations.
Benefits of Procuring Through Google Cloud Marketplace
By listing on Google Cloud Marketplace, Veris AI enables customers to:
- Streamline procurement: leverage existing Google Cloud agreements and simplified billing
- Deploy with confidence: integrate seamlessly with Google Cloud infrastructure, including Google Cloud’s Vertex AI platform, Google Kubernetes Engine, and Google Cloud Storage
- Accelerate time to value: reduce procurement cycles and get agents into production faster
- Draw down committed spend: apply Veris AI usage toward existing Google Cloud commitments
Getting Started
Veris AI is available now on Google Cloud Marketplace. To learn more about how Veris AI can help your team ship reliable AI agents to production, visit veris.ai or contact the team at hello@veris.ai.
Link to video: Veris: The training ground for enterprise AI on Google Cloud
Link to one-pager: Veris AI on GCP