Hexaware Technologies (NSE: HEXT) has entered a multi-year partnership with Anthropic that places Claude at the core of its Zerovity engineering platform and AgentVerse agent-build and governance platform. The agreement gives the IT-services company Preferred Partner status in Anthropic’s Claude Partner Network and is intended to support enterprise clients deploying AI in software engineering and IT operations.
Zerovity is Hexaware’s AI-native engineering platform, while AgentVerse is its platform for building and governing agents. The company said AgentVerse contains more than 600 pre-built agents. Under the arrangement, clients can connect development tools to the Zerovity harness, which Hexaware said applies token optimization and software-delivery agents.
The partnership is significant because it connects a model provider with a services firm whose work includes enterprise implementation and operations. Many organizations can access general-purpose AI models, but moving those models into production involves integration with development tools, data sources, identity systems and change-control processes. Hexaware is positioning the combined platforms as a way to turn defined business outcomes into deployment work.
Its stated starting points include technical debt, security vulnerabilities, backlogs, support tickets, process bottlenecks and software-license costs. Those are target use cases rather than results already demonstrated for customers. The company did not name customers, disclose service pricing, or provide performance measurements for the platforms or their agents.
Hexaware said it has more than 1,100 Claude-certified professionals and will expand Claude training across delivery teams. It also said the agreement builds on an existing authorization to resell Claude through a hosted model service. The partnership includes joint solution development, go-to-market work and customer deployments, according to the company.
For platform and engineering leaders, the relevant question is how an agent layer fits within existing delivery controls. An AI system that writes configuration, analyzes a codebase or generates a proposed workflow still requires review paths, access boundaries and a record of what changed. The source does not provide details about AgentVerse’s policy model, audit features, supported tools or how it separates customer data among deployments.
Those details also affect how a client can exit or modify a deployment. Teams need to know which artifacts remain under source control, how models and prompts are versioned, and whether agent behavior can be reproduced during an incident review. These are ordinary software-delivery questions, but their importance grows when a platform connects model output to operational systems.
That missing detail matters as the scope moves from developer assistance to operations. AI can remove repetitive analysis and generation work, but it can also speed up the propagation of errors if generated output is not validated. Enterprises will need to determine which tasks remain advisory, which can be automated, and where human approval is required before a production change.
The material change is the integration of Claude into platforms designed for engineering and agent governance rather than a standalone assistant deployment. If Hexaware can provide the implementation controls it describes, clients may be able to adopt AI services through a familiar systems-integration model. The eventual operational value will depend on the platform’s governance features and on the specific workflows each organization chooses to place under AI assistance.