Sovos adds Sovi AI tools across tax compliance cloud

Sovos has expanded its Sovi AI tools across the Sovos Tax Compliance Cloud. Adding new capabilities for tax guidance, automation and integration with external AI agents.

At the centre of the update is Ask Sovi, a new assistant built into the tax compliance platform. It is designed to answer regulatory questions during day-to-day work and help users navigate product workflows and documentation.

Ask Sovi draws on a knowledge base built from more than 40 years of tax and regulatory work, with input from more than 100 global regulatory specialists. Sovos says its compliance coverage spans more than 150 countries, and that the assistant can support conversations in multiple languages while interpreting tax and regulatory content across them.

The release also introduces AI features intended to reduce manual work in compliance processes, including tools to diagnose errors, suggest fixes and support remediation after user approval.

Another element focuses on recommendations tailored to a customer's tax registrations, filing history and entity structures. The platform is also being used to flag compliance issues before filing deadlines and before users actively search for a problem.

Alongside these changes, Sovos outlined work on what it calls open agent interoperability. This includes support for Model Context Protocol and Agent-to-Agent frameworks, as well as infrastructure branded as the Sovos MCP Gateway, intended to give AI agents standardised access to the tax compliance platform.

The approach reflects a broader shift in business software, as suppliers move from chat-style assistants towards systems that can take or prepare actions within regulated workflows. In tax, that shift raises questions about oversight, traceability and the use of sensitive customer data.

Sovos says all Sovi AI functions operate within the security perimeter of its Tax Compliance Cloud. Customer data is not used to train external models, and AI-assisted workflows are designed with human oversight and auditability.

"At Sovos, we believe AI should do more than just generate answers from user documentation. It should help tax teams work faster, make better decisions and operate with greater confidence," said Kevin Akeroyd, CEO of Sovos. "With these new Sovi AI capabilities, we are embedding agentic AI across the tax compliance workflow in a way that is practical, governed and purpose-built for the complexity of global compliance," Akeroyd added.

He added that the company is embedding what it describes as agentic AI into tax compliance processes while keeping the system practical and governed for a heavily regulated area of work.

Tax focus

Tax compliance has emerged as a testing ground for applied AI because the work is rules-based, document-heavy and sensitive to jurisdictional changes. Suppliers across the sector have been positioning AI as a way to reduce repetitive tasks while keeping final accountability with finance and tax teams.

For multinational businesses, the appeal lies in managing a patchwork of indirect taxes, e-invoicing rules, reporting obligations and filing deadlines across many markets. A system that can point users to the relevant rule, identify an anomaly or prepare a correction could reduce the administrative burden, although sign-off still rests with the taxpayer.

One part of the new offering is regulatory intelligence that is continuously updated by specialists and used to assess how rule changes may affect each customer's configuration. Sovos also says every recommendation, approval and action is documented for audit visibility.

The company linked the new functions to an existing data layer called Sovos Intelligence, which it says gives customers visibility into invoicing data, filing patterns and emerging risks. That suggests the latest additions are designed to sit on top of information already gathered across the platform rather than operate as a standalone chatbot.

Agent links

The interoperability work may be notable for larger businesses building internal AI systems across finance, legal and operations. By using common frameworks such as MCP and A2A, software providers are trying to make it easier for one AI system to pass requests and context to another without requiring custom integrations for every task.

In practical terms, that could allow tax compliance data or workflows to be accessed by a customer's internal AI tools, subject to controls. Sovos did not provide a timetable for broader orchestration between agents, but says it is building the foundation for that model.

That emphasis on governed access is important in a field where errors can lead to penalties, delayed filings or reporting failures. The company's description of human-approved actions suggests it is trying to draw a line between AI support and unsupervised decision-making.

"Tax compliance is not a category where AI can be useful without governance, traceability and deep domain context," said Swati Garodia, Chief Product Officer at Sovos.

"These capabilities are designed to help customers move faster while keeping humans in control. That combination of intelligence, workflow integration and governance is what makes Sovi AI different," Garodia said.

The new Sovi AI functions are available now or are being rolled out across relevant parts of the platform. Sovos says it serves more than 100,000 customers and processes nearly 20 billion transactions a year across tax compliance activity in more than 150 countries.