Overview
Broadcom is a technology leader that designs, develops, and supplies semiconductors and infrastructure software for global organizations' complex, mission-critical needs. Broadcom combines long-term R&D investment with superb execution to deliver the best technology, at scale. Today, Broadcom’s product portfolio spans advanced semiconductor technologies and infrastructure software solutions.
To support this broad ecosystem, Broadcom partnered with Vectara to modernize customer engagement, unify AI-powered search and conversational experiences, and establish a scalable enterprise-wide agentic AI platform across cloud and on-premise environments.
Getting agents to operate and scale correctly in production involves combining multiple complex components in the right way. Vectara has proven the platform can do this well especially when multimodal data is involved. This drives measurable outcomes in Broadcom’s customer experience and silicon engineering.
Broadcom is leveraging Vectara’s agentic AI platform across two mission-critical areas: (1) the modernization of customer support, and (2) specialized semiconductor engineering assistants.
Vectara proudly helped Broadcom establish their Agentic AI architecture, and we are excited to continue the partnership as they further evolve their usage of AI to drive their business forward.
Accurate, multimodal, enterprise-grade agents across virtual private cloud and on-premise
Broadcom needed to modernize and unify customer support and search experiences across a rapidly expanding portfolio of products, documentation, and engagement channels. Existing keyword-based search and legacy chatbot experiences created friction for some customers, limited self-service success, and increased reliance on live support agents.
Internally, the organization faced growing complexity from fragmented AI and search tooling across business units and environments. Multiple experiments across the company led to a variety of Agentic AI implementation approaches (e.g. “Agent Sprawl”), inconsistent governance, and elevated security and reliability risks.
Furthermore, Broadcom has detailed in their Private Cloud Outlook 2026 publication that there is an observable shift in AI deployments to private clouds. As an industry leader Broadcom required a secure AI platform capable of supporting strict entitlement-based access controls and enterprise governance, while operating fully on-premise.
A unified agentic platform
For the first use case, Broadcom initially deployed Vectara as its unified agent platform to power conversational AI, intelligent search, and enterprise knowledge experiences. Vectara’s platform enables customers and partners to interact naturally with Broadcom’s support ecosystem, orchestrating actions across multiple systems such as password updates, automated ticket creation, and multi-step reasoning.
To support scale and operational efficiency, Broadcom leveraged Vectara alongside Google Gemini models for conversation generation, complex workflow execution, and comprehensive interaction analytics. The platform provides entitlement-aware access controls across every data access path, ensuring users only receive content relevant to the products and services they are authorized to access.
Extending agentic AI to internal use cases
For the second use case and following the success of its customer-facing deployments, Broadcom expanded Vectara into internal use cases.
Broadcom’s semiconductor businesses adopted Vectara's platform, deployed fully on-premise on Broadcom's own private cloud infrastructure.
Running Vectara on VMware Cloud Foundation infrastructure reflects Broadcom's own private AI thesis: the business can benefit from enterprise-class AI running internally, with all required governance, entitlement-aware access control, and hallucination mitigation built in.
Results and business impact
Broadcom successfully consolidated its AI tooling and standardized on Vectara as its enterprise Agentic AI platform. This eliminated Agent sprawl, enabled consistent security and governance controls across environments, and shortened time to production for the highly accurate agents required by the business.
Key operational outcomes include:
- Support scalability
Broadcom achieved a reported 3x containment rate improvement, dramatically improving self-service success and reducing the need for escalation to live support agents.1
- Engineering productivity
Faster debug cycle time and greater than 30% workflow acceleration.2
- Deflection and retrieval
The solution drove up to 45% deflection in internal support and delivered 50% faster internal knowledge retrieval.3
- Trust and grounding
Vectara's hallucination detection and correction keep hallucination rates below 1%.4
- Seamless on-premise alignment
Vectara’s ability to run airgapped and its integration with VMware VKS and Tanzu validates Broadcom's vision of secure, scalable, on-premise production AI.5
Building on this foundation, Broadcom plans to expand their partnership with Vectara deploying AI agents across additional business units and enterprise-wide knowledge workflows.
- 1Production results from Broadcom’s support portal deflection rates by moving to Vectara.
- 2Results confirmed from Broadcom pilot to production.
- 3Results from Broadcom Vectara pilot in chatbot evaluation on multimodal data.
- 4Results from Vectara on-premise testing at Broadcom.
- 5Blog on VMWare VCF/VKS architecture blueprint https://www.vectara.com/blog/take-vmware-private-ai-foundation-with-nvidia-further-with-vectaras-enterprise-agent-platform
