AvePoint launches platform for AI governance & cost

AvePoint has launched the AvePoint Trust Platform for AI governance, security and cost management, aimed at organisations using AI agents across multiple software environments.

The launch marks a broader push by AvePoint to position itself around what it calls AI trust, as companies grapple with questions over what AI agents can access, what actions they can take and how systems can be restored if errors occur.

The platform is designed to give businesses a single place to discover AI agents, models, tools and MCP servers, apply policies to them and generate audit evidence. It also includes budgeting and usage features intended to help companies track AI spending and assess whether deployments are delivering value.

The product supports environments including Microsoft, Google, Salesforce and Anthropic, and extends visibility to Windows and server-hosted agents.

The announcement comes as businesses face a growing mismatch between confidence in AI controls and actual incidents. AvePoint said its State of AI 2026 report found that 88.4% of organisations surveyed experienced at least one AI agent-related security incident in the previous 12 months.

Four areas

AvePoint has structured the platform around four areas: observe, protect, govern and optimise. In practice, that means discovering sanctioned and unsanctioned agents, identifying content that may be overshared or stale, restoring agents and related configurations, assigning ownership and policy controls, and tracking token use and budget consumption.

Backup and recovery features cover Microsoft 365, Power Platform, SaaS applications including ServiceNow, and cloud infrastructure across Azure, AWS and Google Cloud. AvePoint said a Copilot Studio agent, including configuration, prompts, topics and flows, can be restored to a known-good state.

Each agent can also be assigned an owner, policy and data sensitivity context, with attestation and recertification delivered through Slack, Teams and email from discovery through to retirement.

The platform is built on what AvePoint calls a Governance Graph, linking regulatory frameworks including the EU AI Act, ISO 42001 and the NIST AI Risk Management Framework with an organisation's internal policies, ownership structures and information barriers.

Security pressure

Companies have been under pressure to impose more formal controls as AI agents move from experimentation to wider operational use. Unlike traditional software tools, agents can search, read and act across large volumes of enterprise data at speed, creating governance questions tied as much to internal permissions and data hygiene as to external cyber threats.

New governance and security features in the platform include AI inventory, an AI dependency graph and shadow AI detection. These functions are intended to help security, IT and compliance teams identify agents, models, tools and other AI assets, map their relationships and detect unsanctioned activity.

Other additions include monitoring policies to identify ownerless agents and track access to sensitive content. AvePoint also said its data security and classification tooling has been expanded to give businesses more precision in identifying risk, tagging AI-generated content and linking those findings to automated remediation.

Cost focus

AvePoint is also targeting a financial concern that has grown alongside AI roll-outs: visibility into spend. Its AI cost optimisation features bring together spend, usage and adoption data, while new budget request tools are designed to replace email- and spreadsheet-based approval processes.

Managers will be able to review requests with supporting information on usage and return on investment. AvePoint has also introduced adoption coaching and prompts intended to help employees make more effective use of licensed AI tools where utilisation is low.

A further part of the launch is support for Anthropic models and MCP servers. The platform can discover and inventory agents built with Anthropic's Claude and connected MCP servers, alongside agents in Microsoft Copilot Studio and Foundry, Google Vertex AI and Salesforce Agentforce, according to AvePoint.

Dr Tianyi Jiang, Chief Executive Officer and Co-Founder of AvePoint, set out the company's view of the market shift. "Every organization now has access to the same AI capabilities. What truly sets them apart is how far that AI can safely reach, what permissions it actually has, and whether there's a way to reverse course when it makes a wrong call. That's the real test of AI Trust, and it's the foundation for this new agent-driven era," Jiang said.

He added: "For 25 years, we've been understanding how enterprise data behaves in the real world, not just in theory. The AvePoint Trust Platform brings that understanding to every agent, model, and partner that touches that data. It operates independently across vendors and technologies, because in today's world, risk doesn't stop at the vendor's edge."

Randy Simmons, Senior Managing Director, EY AI Workplace Practise, said large companies were looking for stronger oversight as AI deployments expand. "As organizations scale AI, they need the right foundation to innovate with trust and without introducing new risks," Simmons said.

"AvePoint addresses a critical need we are seeing in the large enterprise market, giving organizations more visibility, control and protection as their AI environments grow. We believe that together, EY and AvePoint can help our clients rapidly scale AI with greater trust and confidence," he added.