Summary
- Vanta has introduced expanded AI governance, risk quantification, autonomous vendor assessment and agent-assisted compliance capabilities at VantaCon 2026.
- The platform aims to connect security signals, supplier information and compliance controls with assessments of operational and financial exposure.
- AI Governance by Vanta is scheduled for paid availability in January 2027, with adversarial testing and runtime enforcement planned for later in the year.
Vanta is expanding its enterprise risk platform to connect AI governance, supplier assessments and financial exposure, bringing automated compliance workflows closer to the security information generated by systems operating across a business.
The San Francisco company introduced AI Governance by Vanta, Risk Quantification, Autonomous Vendor Assessment and additional functions for its Vanta Agent at VantaCon 2026 on 12 November. The developments are intended to give security and compliance teams a more consistent view of the systems, suppliers and risks they are responsible for overseeing.
Vanta serves more than 20,000 customers, according to its latest company figures. Its existing platform brings together information used for security monitoring, compliance evidence and third-party risk management, while the latest additions extend automation into assessments that have traditionally required substantial manual review.
Availability varies across the announced capabilities. Although Vanta is introducing expanded risk and compliance functionality within its existing platform, its full AI Governance offering is scheduled to become a paid product in January 2027. Additional capabilities for testing and controlling AI agents are expected later, rather than being available immediately to every customer.
Connecting security evidence with business exposure
Governance, risk and compliance software commonly helps organisations maintain policies, record assessments and demonstrate whether required controls are operating. Those functions support audits and internal oversight, but the resulting evidence does not always explain how a particular weakness could affect business operations.
A failed technical control might expose a relatively unimportant system, while an apparently similar failure involving customer information or a critical supplier could have much greater consequences. Understanding that difference requires information about the assets involved, how they are used and which business processes depend on them.
Vanta's Trust Graph is intended to connect such relationships across systems, people, suppliers, risks and compliance obligations. The company wants its risk management functions to use that context
when identifying exposures and directing attention towards issues with potentially significant consequences.
Its Risk Quantification capability adds a financial dimension, translating selected security and operational risks into estimates of possible monetary impact. The approach could help organisations compare competing demands for investment, particularly when technical weaknesses are described through measures that are difficult for finance and operational teams to evaluate.
Financial estimates depend, however, on assumptions about the likelihood and consequences of an event. Calculations may need to consider service interruption, recovery costs, contractual obligations and potential liabilities, alongside the organisation's existing safeguards.
An estimate generated from incomplete information can give a misleading impression of precision. A calculated figure is not an independently established loss, and the company has not published evidence demonstrating that its new estimates consistently predict the financial consequences of real incidents.
Organisations evaluating the functionality will therefore need to understand the underlying data, calculation methods and opportunities for human review. Its value will depend on whether reported financial exposure supports better decisions rather than simply adding another measure to an existing risk dashboard.
Supplier assessments move towards automated workflows
Vanta is also extending its use of automation across supplier risk management, where organisations frequently have to collect documentation, examine security questionnaires and determine whether external providers satisfy contractual or regulatory requirements.
These processes can involve multiple teams and substantial correspondence. Suppliers may be required to provide evidence of security controls, explain how they handle sensitive information and demonstrate the arrangements they use to manage operational disruption.
Autonomous Vendor Assessment is intended to help gather and examine this information, identify outstanding questions and connect relevant findings to an organisation's broader risk records. Vanta's existing agent-assisted capabilities can also support supplier intake and follow-up activities.
Automating the collection and preliminary evaluation of evidence may reduce repetitive administrative work. Nevertheless, documents provided by a supplier do not necessarily establish that its controls function as described, particularly when circumstances have changed since the material was prepared.
Assessment also depends on the supplier's role. A provider holding sensitive customer records may require different scrutiny from one supplying a non-critical administrative service, even where both submit similar security documentation.
By connecting supplier assessments to information about business systems and risks, Vanta aims to reflect those differences within its automated processes. The reliability of the results will depend on accurate supplier records, suitable assessment criteria and the ability to escalate findings requiring human judgement.
The company also wants its Vanta Agent to undertake longer sequences of compliance work, involving employees when decisions or approvals are needed. Planned integrations with workplace applications, including Claude, ChatGPT, Cursor and Slack, would allow some activities to take place through tools already used by employees.
Such integrations create additional access management considerations. An agent capable of examining evidence, changing records or initiating follow-up tasks must operate within permissions appropriate to its responsibilities, particularly where compliance records influence contractual or regulatory decisions.
AI governance introduces a longer delivery timetable
Vanta's November product presentation follows its July introduction of an AI governance programme and its more recent agreement to acquire Pillar Security, a specialist in protecting AI systems during operation.
Techopia previously examined the proposed Pillar acquisition and its implications for AI runtime security. The developments presented at VantaCon are distinct, extending the company's risk and compliance proposition alongside the security technology it intends to integrate.
The published roadmap establishes several delivery stages. AI inventory and runtime monitoring are targeted for January 2027, with adversarial testing expected in spring and runtime enforcement scheduled for summer.
Inventory functions are intended to identify AI applications, models, agents and related infrastructure, including Model Context Protocol servers. Connecting that information to system owners and approved uses could provide a basis for establishing which applications have access to sensitive data or important business functions.
Runtime monitoring would add information about how those systems behave during operation, while adversarial testing could help identify weaknesses that are not apparent from policy reviews or conventional software assessments.
The planned enforcement capability goes further by seeking to prevent unauthorised actions before they are completed. Its effectiveness will depend on the systems supported, the available technical integrations and the rules organisations configure to govern automated behaviour.
These capabilities have not all entered general commercial use, and Vanta has not established their performance across the full range of enterprise AI environments. Customers considering the platform must distinguish existing compliance functions from the security controls still under development.
Compliance responsibility remains with the organisation
European businesses adopting AI face obligations arising from data protection law, sector-specific requirements and the EU AI Act. The applicable duties depend on the type of system, its intended purpose and the role an organisation plays in developing or deploying it.
Software that records AI systems and connects them to policies can support governance, but purchasing a monitoring platform does not independently establish legal compliance. Organisations remain responsible for determining which obligations apply and demonstrating that their controls meet the relevant requirements.
Financial risk estimates and automated supplier reviews likewise require oversight when their conclusions influence significant decisions. Incorrect system classifications, outdated documentation or unsuitable assumptions could affect spending priorities and the treatment of commercial partners.
Vanta's wider approach reflects growing interest in connecting compliance evidence to operational activity rather than maintaining separate records that are updated only for periodic reviews. The company's customers will be able to evaluate that proposition more fully as its planned AI inventory, monitoring and enforcement capabilities become available.
Its published AI governance roadmap identifies January 2027 as the next major delivery milestone. The quality of the eventual integrations, the accuracy of financial risk assessments and the scope of automated actions will determine whether the expanded platform provides more useful oversight of enterprise AI systems.












