Skip to content
← Back to News

product-launch · October 14, 2026

By Joe Jurczyk, Editor

Gupshup launches Avatar Agent, a self-serve video agent builder

The conversational messaging company added video-based AI agents to its messaging and voice line-up, deployable on websites, mobile, tablets and kiosks, with a free branded tier and pay-as-you-go pricing.

conversational agents
video agents
enterprise
product launch

Gupshup has launched Avatar Agent, a self-serve platform for building AI agents that interact with customers over video, extending its existing messaging and voice offerings. Businesses can deploy the agents on websites, mobile devices, tablets and kiosks without an integration process.

The agents can answer queries, explain products, qualify leads, guide users through processes, host product walkthroughs and webinars, and hand off to a human. Layouts, visual styles, brand elements and messaging are customisable, with industry presets available. Agents can be configured through Gupshup SuperAgent, ChatGPT, Claude and other AI assistants using plain-language descriptions. The product builds on Gupshup's recent Voice AI release, and existing messaging agents can be extended to voice and video. Target sectors named are banking and financial services, e-commerce, real estate, travel and hospitality, and education. Pricing is a free tier carrying Gupshup branding, with pay-as-you-go pricing above it.

Founder and CEO Beerud Sheth said avatar agents "bring back the personal touch."

Why it matters for digital humans: The notable thing here is the channel ladder rather than the avatars. Gupshup reached video by stacking it on an installed base of messaging and voice agents, which is a different go-to-market from the avatar-first vendors — the customer already has the intent model and the escalation path, and video is a presentation layer bolted on top. A free tier with vendor branding also pushes interactive avatars toward commodity pricing, which is where chat widgets ended up. The announcement states no latency figures, no avatar technology partner, no supported language count and no adoption numbers, so there is nothing here to assess conversational quality against; treat the capability list as vendor-described.

Related research