8×8 has launched AI Studio, a tool that lets contact center teams build and deploy AI agents on the platform they already operate – without specialist developers, professional services engagements, or months of integration work.
Available now in early availability for existing customers, the product is designed to address one of the more persistent frustrations in enterprise AI adoption: the gap between deciding to deploy an agent and actually getting one into production.
Indeed, according to the Metrigy Customer Experience Optimisation 2025-26 report, nearly three in four CX leaders would rather build their own AI agents than purchase off-the-shelf solutions, with trust and domain expertise cited as the main reasons.
That appetite has existed for a while, but the practical barriers have kept many organizations stuck at the planning stage.
Hunter Middleton, Chief Product Officer at 8×8, made the case for why native infrastructure changes the equation:
“8×8 AI Studio is not an AI layer sitting on top of a communications platform, it’s AI embedded in the infrastructure itself. The LLM has direct access to real-time voice data, network telemetry, and the full interaction context that external tools typically cannot access.”
“That direct access also eliminates the transcription intermediaries responsible for the latency and drop-offs that destroy the natural conversation experience on older architectures.
“That is what separates AI that demos well from AI that holds up at production scale.”
That last distinction is of particular interest. A lot of AI agents in the contact center space perform impressively in controlled demos but struggle once they hit real conversation volume.
Transcription lag, missing context, and broken handoffs are just some of the familiar failure points for CX teams who have already been down the vendor integration road.
Why Native Infrastructure Is the Differentiator
The centerpiece of AI Studio is a Builder tool that allows any user to describe what they need in plain language and have the agent created, tested, and ready for deployment across voice and digital channels in minutes.
According to 8×8, dozens of customers across more than 15 verticals are already running hundreds of agents during the early availability period, covering inbound routing, outbound follow-up, sales qualification, scheduling, and internal helpdesk triage.
Sheila McGee-Smith, President and Principal Analyst at McGee-Smith Analytics, offered context on why the native approach resonates with a specific segment of the market:
“A business user can describe what they want an agent to do in natural language, and the native Builder creates it, including the ability to take real action such as scheduling appointments or filing tickets.
“This combination of accessibility for non-technical teams and the flexibility to expand into agentic workflows on a single platform is what will allow companies to transform AI from a concept to repeatable outcomes.”
What Teams Are Actually Building
On the use case side, 8×8 is positioning the tool to handle more than standard self-service flows.
Early deployments include always-on inbound coverage with identity verification, proactive outbound agents for appointment confirmation and open service requests, sales qualification workflows with live Salesforce integration, and employee-facing helpdesk triage.
There is also a personal productivity angle, with individual 8×8 users able to configure their own agent through the Builder to screen calls and manage after-hours interactions without IT involvement.
From a commercial standpoint, the entry point is accessible. During early availability, AI Studio is included at no extra licensing cost for existing customers and carries a free tier for building and testing. Consumption fees kick in once agents are live in production.
Whether AI Studio ultimately moves the needle on the AI execution gap will depend on how well those production agents hold up across the messier, higher-volume scenarios that contact centers deal with daily.
But for organizations already operating on the 8×8 platform, the case for at least testing seems a bit of a no-brainer: the infrastructure is already there, and the starting cost is zero.
