What are Cisco AI PODs?

Cisco AI PODs for Collaboration bring AI-powered real-time transcription to on-premises Webex environments. The product supports Cisco Meeting Server and is designed specifically for organizations that cannot or will not move collaboration workloads to the cloud. All AI processing runs locally, with no data leaving the customer's defined infrastructure boundary.

Which organizations is this product built for?

The primary targets are organizations with strict data sovereignty, compliance, or security requirements: defense contractors, financial institutions, healthcare providers, and public sector entities. These are segments that have been excluded from using AI-powered collaboration features because those features have historically required cloud connectivity. Air-gapped network deployments are explicitly supported, with no external connectivity required.

What does the product actually do, and what are its current technical limits?

Each Collaboration AI POD delivers real-time transcription through Automatic Speech Recognition, producing live closed captions and transcription output. The system supports five languages: English, French, Spanish, Italian, and German. A single POD handles 700 to 800 concurrent transcription sessions. Multiple PODs can be deployed together with built-in load balancing using a least-load, top-down distribution method, and the architecture includes fault tolerance without requiring clustering. Encryption is handled via TLS 1.3, with API key authentication and encrypted storage throughout.

How does the POD architecture scale, and what does that mean for a multi-room deployment?

The multi-POD architecture scales horizontally: additional appliances are added and the SDK distributes sessions automatically across them without manual intervention. At 700 to 800 concurrent sessions per POD, a single appliance covers a substantial simultaneous meeting load. An organization running 1,500 concurrent transcription sessions, for example, would require at least two PODs.

What is on the product roadmap beyond these initial transcription features?

Cisco has identified AI Assistant Summarization, additional language support, and coverage for on-premises Calling and Contact Center as the next capability additions. The company also indicated that cloud-based Webex customers with privacy requirements will be able to use AI PODs in hybrid configurations, bringing local AI processing to deployments that are otherwise cloud-connected.

How does this fit into Cisco's broader on-premises and cloud strategy for Webex?

Cisco has framed the AI POD as one deployment model within a single unified platform, positioning the same AI capabilities across cloud, hybrid, and on-premises configurations. The company's stated strategic direction, which it calls Connected Intelligence, describes a continuous loop of human-to-AI-to-human interaction intended to work regardless of where the infrastructure sits.

Corporate · June 25, 2026

Cisco AI PODs for Collaboration: What On-Premises AI Transcription Means for Security-Sensitive AV Deployments

Cisco brings AI-powered transcription to air-gapped and on-premises Webex environments, with no cloud dependency required.

By AVinBrief Editorial Desk · June 25, 2026

What are Cisco AI PODs?: Cisco AI PODs for Collaboration bring AI-powered real-time transcription to on-premises Webex environments. The product supports Cisco Meeting Server and is designed specifically for organizations that cannot or will not move collaboration workloads to the cloud. All AI processing runs locally, with no data leaving the customer's defined infrastructure boundary.
Which organizations is this product built for?: The primary targets are organizations with strict data sovereignty, compliance, or security requirements: defense contractors, financial institutions, healthcare providers, and public sector entities. These are segments that have been excluded from using AI-powered collaboration features because those features have historically required cloud connectivity. Air-gapped network deployments are explicitly supported, with no external connectivity required.
What does the product actually do, and what are its current technical limits?: Each Collaboration AI POD delivers real-time transcription through Automatic Speech Recognition, producing live closed captions and transcription output. The system supports five languages: English, French, Spanish, Italian, and German. A single POD handles 700 to 800 concurrent transcription sessions. Multiple PODs can be deployed together with built-in load balancing using a least-load, top-down distribution method, and the architecture includes fault tolerance without requiring clustering. Encryption is handled via TLS 1.3, with API key authentication and encrypted storage throughout.
How does the POD architecture scale, and what does that mean for a multi-room deployment?: The multi-POD architecture scales horizontally: additional appliances are added and the SDK distributes sessions automatically across them without manual intervention. At 700 to 800 concurrent sessions per POD, a single appliance covers a substantial simultaneous meeting load. An organization running 1,500 concurrent transcription sessions, for example, would require at least two PODs.
What is on the product roadmap beyond these initial transcription features?: Cisco has identified AI Assistant Summarization, additional language support, and coverage for on-premises Calling and Contact Center as the next capability additions. The company also indicated that cloud-based Webex customers with privacy requirements will be able to use AI PODs in hybrid configurations, bringing local AI processing to deployments that are otherwise cloud-connected.
How does this fit into Cisco's broader on-premises and cloud strategy for Webex?: Cisco has framed the AI POD as one deployment model within a single unified platform, positioning the same AI capabilities across cloud, hybrid, and on-premises configurations. The company's stated strategic direction, which it calls Connected Intelligence, describes a continuous loop of human-to-AI-to-human interaction intended to work regardless of where the infrastructure sits.

The Take

What does the Cisco AI POD release mean for integrators specifying collaboration systems in regulated or security-sensitive environments?

For years, the practical answer to “can we get AI transcription in our on-premises Webex deployment?” has been no. That changes the specification calculus for integrators working in government, healthcare, and financial services, where data sovereignty requirements have been a hard stop on AI-enhanced UC features.

The more immediate question integrators should be asking: how does the POD get sized and specified for a given deployment? The 700-to-800-session figure per appliance is a useful planning anchor, but real-world session loads in large enterprise or agency environments are rarely uniform. Peak capacity during all-hands meetings or coordinated operational events can spike well above average. Integrators should clarify with Cisco how load balancing behaves at or near POD capacity, what the failover behavior looks like during a POD outage, and whether any on-site infrastructure prerequisites (like rack space, power, network segmentation, etc) affect the deployment scope.

The five-language support covers the primary Western European languages, but multilingual organizations operating in, say, Arabic, Mandarin, or Japanese should be watching the roadmap closely.

One structural consideration worth tracking: the AI POD lives at the intersection of AV systems integration and IT infrastructure. The commissioning scope spans Cisco Meeting Server configuration, network segmentation for air-gapped environments, and physical appliance installation. Whether the specification, system design, and installation work lands with an AV integrator, the IT team, or a Cisco networking partner may depend on how contracts are structured. Integrators in regulated-sector accounts should always clarify scope of work before the proposal stage.

Sources

Cisco AI PODs for Collaboration Are Now Generally Available: Delivering Secure, On-Premises AI, Cisco Webex Blog (press release) · Authored by Ashish Chotai, Director of Product Management, Cisco Collaboration Technology Group.