Navan has deployed Nvidia's Nemotron 3.5 Super VL, a vision-language model, to power its AI travel agents, according to StreetInsider.

The deployment names two specific players: Navan, the corporate travel and expense platform, and Nvidia, whose Nemotron model family now supplies the underlying intelligence for Navan's automated agents. Neither company disclosed deal value, deployment scale, nor the share of bookings the AI agents now handle.

What does the deployment change?

The announcement confirms Navan is building its travel-agent experience on a frontier third-party model rather than a proprietary stack. For a corporate travel platform, that choice affects speed of iteration and cost per interaction — the two variables that determine whether automated servicing scales profitably.

Nemotron 3.5 Super VL is a vision-language model, meaning it can process both text and images. In a travel context, that capability applies to documents such as boarding passes, invoices and receipts — the unstructured inputs that drive back-office cost in corporate travel servicing.

What remains unverified?

StreetInsider's report carries no measured results. There are no figures on:

• Adoption: how many Navan customers use the AI agents
• Volume: what share of bookings or service requests the agents handle
• Economics: cost savings or pricing changes tied to the deployment
• Timing: when the model entered production

Sellers and buyers of corporate travel technology should treat the deployment as a capability claim, not a measured outcome. The revenue and distribution consequences depend on figures Navan has not yet published.

Navan has not said whether it will report adoption metrics for the Nvidia-powered agents; any distribution shift in corporate travel booking flows will show up in those numbers first.