SoundHound AI Reimagines AI-Human Contact Center Handoffs

SoundHound has added a new "Human Assisted Resolution (HAR)" capability to its conversational AI platform: SoundHound AI OASYS.

The new feature introduces a fresh approach to AI-human handoffs in the contact center.

Traditionally, the AI agent takes a customer conversation as far as it can, then escalates to a human agent when it isn’t sure of an answer, needs approval for an action, or a customer demands it. The human then completes the conversation.

Yet, HAR lets an AI agent tap a human agent to complete only a specific part of the resolution flow. 

The AI agent still leads the conversation, even as the human agent fulfills the task in the background, and - after they’ve done so - moves the interaction forward.

For example, consider a flow that requires explicit human approval or a judgement call. The AI can nudge a human behind the scenes to make the call, before progressing the conversation.

The result? Organizations can orchestrate more AI-led experiences end-to-end, rather than stopping short at the point when a human must take over.

Meanwhile, the customer doesn’t feel like they’re being passed from pillar to post. The AI agent leads the conversation while the human only gets involved behind the scenes.

SoundHound AI shared a video example of how this works on LinkedIn, as shown below. 

“We believe this is how human-in-the-loop is meant to operate in the agentic era,” added Jack Gantt, Director of Product Marketing at SoundHound AI in a company blog post. 

“With HAR, a human never has to take the conversation over… They contribute one piece of judgment, in seconds, and move on to the next."

Jack Gantt

Director of Product Marketing at SoundHound AI

In this sense, the human acts like another task agent within an agentic resolution flow.

That’s illuminating for those reconsidering the contact center workforce of tomorrow. After all, humans won’t just be picking up complex, messy contacts, but potentially performing specific tasks within resolution flows that AI agents cannot yet do or are not allowed to do. 

It's also fascinating to consider, in a future of machine customers, how these handovers may become necessary for contact centers coming to terms with agentic to agentic.

Nevertheless, before we get too excited, other providers must play catch-up on their human-AI handoffs before this becomes a widespread possibility. 

Not All Human-AI Handoffs Are the Same

Even before HAR, many providers trailed SoundHound in human-AI handoff capability. 

For instance, SoundHound lets a human agent step into the same conversation window as the AI agent and take over. There’s no formal escalation required.

That’s still a capability many providers cannot offer. 

Of course, most will promise a “warm handover”, which summarizes the context of the AI-led interaction and shares it with the human agent taking over. Yet, beyond that, innovation has proven slow. 

That’s evident in a recent Forrester evaluation of 13 conversational AI providers (specifically for employee service, not customer service). 

In the study, only two providers - SoundHound and EasyVista - scored a perfect five out of five for their handoff. Conversely, several prominent providers, including Microsoft and IBM, scored one out of five.

Despite this, some other providers in the space are pushing boundaries.

For instance, Retell AI introduced an “agentic warm transfer”, allowing human agents to ask questions to the customer-facing AI agent to gain further context into the customer contact before accepting the escalation. 

Nevertheless, SoundHound AI has seemingly found another strong - although likely short-term -  differentiator here, with another being its multimodal AI. 

Also at SoundHound… The LivePerson Acquisition Is Complete

As SoundHound’s team keeps innovating, it is also set to onboard another rival platform after its $43 million LivePerson acquisition went through earlier this month. 

Notably, it’s the third conversational AI player, chiefly serving contact centers, that the company has snapped up since 2024, following Amelia and Interactions. 

The main motive is likely to capture LivePerson’s long-tenured enterprise relationships, with the vendor reportedly still working with 25 Fortune 100 companies.

When announcing the move, SoundHound also pointed to the strength of LivePerson's messaging channel infrastructure and how it would complement SoundHound's traditional strengths in voice AI within a single platform.

Moreover, LivePerson has a global team of contact center experts, experienced in tapping into the legacy infrastructures many large enterprises still have.

However, LivePerson - which once had a market cap of approximately $4.5 billion - was a company in decline. 

Given this, SoundHound is likely to be buying LivePerson primarily for its scale and relationships, not its technology. 

With 650+ conversational AI providers selling to customer service and support teams, further market consolidation is likely.