Avalara has launched Versori by Avalara, an agentic integration fabric designed to help developers connect enterprise systems with its tax and compliance services. The platform uses specialized AI agents to analyze APIs and schemas, generate integration logic, test connections and deploy governed connectors, giving enterprises another approach to integrating AI-driven compliance into ERP, billing, e-commerce and custom software environments.
Enterprise software integration has traditionally required developers to study API documentation, map data models, write transformation logic, build authentication workflows and test connections across multiple systems. Avalara is now applying agentic AI to that process with Versori by Avalara, a developer platform designed to automate much of the work involved in connecting business applications to tax and compliance services.
The product is available through Avalara’s Developer Portal and is aimed primarily at developers and technology partners building integrations around Avalara’s tax and compliance platform. According to Avalara, specialized AI agents can move an integration from initial requirements through testing and deployment rather than simply generating snippets of code.
That distinction is important as enterprises increasingly experiment with AI agents that can perform multi-step software-development tasks. Instead of treating generative AI as a coding assistant that responds to individual prompts, agentic integration platforms attempt to give AI systems responsibility for a defined workflow, including analyzing technical inputs, making mapping decisions and executing tests.
Four-agent approach to integration
Versori organizes the process around four functions: Plan, Connect, Build and Deploy.
The Plan agent analyzes endpoint schemas, API contracts and compliance requirements to establish an integration architecture. Connect handles authentication and system-to-system communication, while Build generates transformation logic, mapping rules and error-handling code. Deploy then automates testing, validates the connector against Avalara’s certification requirements and publishes the integration.
Avalara’s current developer documentation describes a similar workflow in which developers provide an OpenAPI specification, GraphQL endpoint, database schema or sample CSV/JSON data. The AI agents can then map fields to Avalara’s data model, generate integration logic and run test transactions against the tax calculation engine before developers review and deploy the result.
The platform also gives customers and partners access to the generated connector code and business logic, according to Avalara. That is a significant consideration for enterprise AI adoption because organizations often need visibility into how automated systems transform business data and make integration decisions.
From code generation to governed workflows
The move reflects a broader shift in AI development platforms from standalone code generation toward workflow-oriented agents.
Generic coding assistants can generate functions or suggest implementation approaches, but enterprise integration involves more than writing code. Developers must understand source and destination schemas, authentication requirements, data transformations, error handling, testing and operational controls.
Avalara is effectively packaging those steps into an agentic workflow around its own APIs and compliance requirements. Its developer platform also supports conventional REST APIs and SDKs alongside newer agent-based approaches, including Model Context Protocol connectivity.
That hybrid model is relevant for enterprises that are not ready to hand an entire integration workflow to autonomous software. Teams can continue using conventional APIs while introducing AI agents where they provide practical development assistance.
Tax compliance creates a demanding AI use case
Tax and compliance software is a particularly consequential environment for agentic automation because integrations ultimately feed financial transactions and regulatory processes.
Avalara says its broader platform supports tax and compliance operations across more than 190 countries and connects with enterprise systems including ERPs, e-commerce platforms, marketplaces and billing applications. The company also says its tax platform processes more than 54 billion transactions annually. Those figures are company-reported rather than independent market measurements.
The integration layer therefore becomes an important part of Avalara’s wider AI strategy. Its platform already uses AI agents for tax and compliance workflows, while MCP servers allow external AI agents to discover and interact with Avalara services.
This creates a broader architecture in which AI does not necessarily replace the underlying enterprise system. Instead, agents operate as an orchestration layer connecting existing software, APIs and specialized services.
The governance question remains
The potential benefit of agentic integration is speed, but enterprise adoption depends on more than how quickly an agent can produce code.
Avalara’s own July 2026 survey of more than 1,500 finance leaders found that 44% were only somewhat confident they could explain an AI agent’s actions to an auditor or regulator, while just 7% said their organizations prioritized governance over deployment speed. Because the research was commissioned and published by Avalara, it should be treated as company-sponsored survey data rather than an independent industry benchmark.
Versori’s review and audit-trail capabilities address part of that challenge by allowing developers to inspect mapping decisions before deployment. The approach illustrates an emerging enterprise AI pattern: autonomous systems can perform more of the implementation work, while humans remain responsible for review, approval and production deployment.
For the wider AI infrastructure and enterprise software market, Avalara’s launch is another example of agentic AI moving from conversational interfaces into specialized technical workflows. The competitive question will increasingly be whether AI agents can reliably execute complex enterprise processes while maintaining the visibility, testing and controls required for production systems.
Market Landscape
Enterprise AI is shifting from chat-based assistance toward agents that can perform multi-step tasks across applications, APIs and data sources. Integration is a natural target because it combines structured inputs, repeatable workflows and clearly defined technical outcomes.
Avalara is approaching the market from a specialized position: rather than offering a general-purpose integration agent, Versori is designed around connecting enterprise systems to Avalara’s tax and compliance capabilities. The company’s developer platform supports multiple integration paths, including agent-to-agent MCP connections and conventional REST APIs.
The broader market includes integration-platform vendors, API management providers, cloud platforms and AI coding tools. The differentiating factor for agentic integration will likely be how effectively these systems handle enterprise-specific schemas, security, testing, observability, governance and human approval.
Top Insights
- Avalara’s Versori applies specialized AI agents to planning, building, testing and deploying enterprise integrations rather than limiting AI to code suggestions.
- Developers can provide APIs, schemas or sample data and have agents generate mapping and integration logic against Avalara’s systems.
- The platform combines agentic workflows with conventional REST APIs, creating multiple paths for enterprises with different automation requirements.
- Human review remains part of the deployment process, addressing governance concerns surrounding autonomous AI actions in financial workflows.
- Avalara is extending agentic AI beyond tax calculations into the infrastructure layer that connects compliance services with enterprise software.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
