As artificial intelligence becomes a bigger part of how software gets built, organizations face a new challenge: moving quickly without introducing new security and compliance risks.

At Zivian Health, we believe responsible AI adoption requires more than choosing the right tools. It requires building the safeguards around those tools at the same time.

That is why Zivian is working with AI security company HiddenLayer as a design partner for its new Agent Harness Security solution, announced on August 3, 2026. The technology is designed to protect AI-powered software development environments from emerging risks while allowing engineering teams to continue benefiting from AI-assisted development.

Building Safeguards Into AI-Enabled Development

AI coding agents have rapidly evolved beyond tools that simply suggest lines of code. Today, they can interact with repositories, execute commands, install dependencies, edit files, and use information from other systems. Those capabilities can make development dramatically more productive, but they also create new security considerations.

HiddenLayer’s Agent Harness Security operates within these AI coding workflows to monitor agent activity and identify risks such as prompt injection, exposure of sensitive credentials, unsafe commands, and malicious or unexpected software dependencies. It can also redact sensitive information before it reaches an AI model and intervene when an agent encounters potentially compromised information.

For Zivian, those protections address risk in both directions.

They help defend development environments against external threats that could manipulate AI tools or introduce compromised code. At the same time, they create safeguards against ordinary human mistakes, such as inadvertently exposing credentials or sensitive information while working with an AI-enabled development tool.

As Jaclyn Miller, Chief Product Officer, and Vanessa DeGennaro, VP of Engineering at Zivian Health, shared in HiddenLayer’s announcement:

“Our focus is not only on enabling engineers to benefit from AI, but also on empowering citizen developers to experiment with confidence — all while ensuring agent activity can be monitored, governed, and constrained when necessary.”

Responsible AI Requires Infrastructure

Healthcare organizations cannot approach new technology with a choice between innovation and responsibility. They need both.

Zivian builds compliance infrastructure for healthcare organizations managing and scaling NP and PA workforces. That work is grounded in a simple principle: the right infrastructure should make responsible growth easier, not harder.

We are applying that same principle internally as our own technology evolves.

AI is changing how software teams work, and security practices have to evolve alongside it. Putting safeguards around AI-enabled development early allows our teams to experiment, learn, and build faster while maintaining the controls expected of a company operating in healthcare.

For an early-stage technology company, investing in this kind of infrastructure requires intention. We believe it is worth making that investment now rather than treating security and governance as problems to solve later.

We believe that companies that pair AI adoption with the right safeguards, governance, and infrastructure will be best positioned to capture its value responsibly at scale.

Read HiddenLayer’s announcement to learn more about Agent Harness Security and Zivian’s participation in its launch.