New Switch Trust and Switch ROI capabilities help businesses assess AI agent risks, measure business outcomes and optimise costs while retaining control over enterprise data and context.
SandboxAQ is expanding its Switch AI platform with new capabilities designed to help enterprises govern, secure and measure the performance of AI agents as their adoption spreads across business operations.
Announced on October 20, 2026, Switch Trust and Switch ROI build on the company’s open-source Switch Rooms offering, which brings employees and AI agents together across workplace platforms while allowing organisations to retain control of shared context, knowledge and interaction history.
The expansion addresses a growing enterprise challenge: AI agents can increasingly make decisions and execute tasks, but organisations often lack consistent ways to assess their security, track their costs and determine whether they deliver measurable business value.
“More agents doesn’t automatically mean more value.” — Mohammed Aboul-Magd, GM of Switch AI, SandboxAQ
Switch Trust provides four capabilities to manage agent risk throughout the lifecycle. Discover inventories agents, models, tools, Model Context Protocol (MCP) servers and dependencies across connected environments and code repositories. Scan identifies security vulnerabilities through static analysis and AI-assisted triage, while Evaluate tests agents against threats such as prompt injection, jailbreaks and data leakage.
Monitor extends oversight into production environments, capturing agent activity and enforcing guardrails against risks including exposure of sensitive information, personally identifiable information and unsafe interactions. Organisations can block, redact, log or flag problematic behaviour, while agents operating within Switch Rooms receive guardrail protection through the platform.
Switch ROI focuses on the economics of agent deployment. It tracks spending and performance at agent, session and interaction levels, linking costs to outcomes such as resolved support tickets, accelerated code development and automated tasks. Its measurement capabilities include cost per outcome and net return on investment, helping businesses compare agent performance against human-led baselines.
The platform also recommends ways to reduce unnecessary spending, including selecting more appropriate models, eliminating unused tools, improving context caching and addressing repeated retries or inefficient agent loops. Recommendations include estimated monthly savings and confidence levels.
SandboxAQ cited industry concerns over the sustainability of agentic AI projects. Gartner has predicted that more than 40% of agentic AI projects could be cancelled by the end of 2027, while the company said only 7% of leaders report having established a return on their AI investments.
Mount Sinai West’s MSW Lab is already using Switch Trust to support G Board, a framework for physicians building clinical applications with AI agents.
“In the MSW Lab, physicians build real healthcare software with AI agents, so we have to prove those agents behave safely,” said Dr. Jeffrey Bander, Chief of Cardiology at Mount Sinai West and Associate Professor of Cardiology and Artificial Intelligence at the Icahn School of Medicine at Mount Sinai. “Switch AI gives us a way to build that trust into how we work, so our physicians can focus on solving clinical problems without having to become AI security experts.”
Mohammed Aboul-Magd, GM of Switch AI at SandboxAQ, highlighted the importance of measuring agent economics beyond adoption figures.
“We’ve seen this across many users: in cases where we thought agents could help avoid a net-new hire, the AI spend quickly became more expensive than hiring a full-time person,” Aboul-Magd said. “That’s why ROI can’t just mean ‘we adopted AI.’ You need to know what every agent costs, what it delivers and how to make it better over time. That’s the problem Switch ROI is built to solve.”
SandboxAQ CEO Jack Hidary said enterprises must apply the same accountability to AI agents as they do to employees, assessing qualifications, costs and results while retaining the ability to change providers without losing accumulated organisational knowledge.
“We’re entering a world where companies will have hundreds, potentially thousands, of AI agents working alongside their employees,” Hidary said. “But you wouldn’t hire a thousand people without knowing their qualifications, what they’ll cost or whether they’re delivering results. The same discipline needs to apply to AI agents. And as the technology advances, companies must be able to switch to better agents and models without losing the knowledge they’ve built. The intelligence belongs inside the company, not the AI provider.”
With Switch Trust and Switch ROI, SandboxAQ is positioning agent governance and performance measurement as essential controls for moving AI from experimentation into production.
