THE SERVICE DESK
SMARTASSIST / THE CONTEXT MUST TRAVELVOICE → DIGITAL → HUMANTHE NEXT CHAPTER: XO v11
Company / AI & Enterprise

SmartAssist.ai: Give the queue an exit.

Kore.ai built SmartAssist to move routine requests out of the phone queue and carry the conversation to a human when needed. Its most useful idea is simple: the customer should never have to start again.

Advertisement

A customer with a simple question can become a complicated problem merely by calling at the wrong hour. The question has not changed. The office has closed. Or the queue has filled. Or the person who finally answers needs the story told again. SmartAssist.ai, the contact center offering from Kore.ai, was built around these small failures of coordination. Its subject is the distance between asking for help and actually getting somewhere.

The useful bits
  • Routine requests can move through voice or chat automation.
  • A human can take over with the conversation’s context.
  • Existing SmartAssist accounts have a documented path to XO v11.

That makes it an interesting business to examine without getting distracted by a talking machine. Speech is the visible trick. Underneath sits a series of decisions: what does this customer want, which system can act, and when should a person step in? A pleasant voice cannot rescue a journey that sends someone to the wrong place.

An exit, hidden in a text message

In Kore.ai’s 2020 materials, SmartAssist.ai was a cloud service for call automation and digital deflection. The proposition was practical. Keep the existing contact center platform. Use business rules to decide which calls should move elsewhere. Send a caller toward a virtual assistant or a digital conversation, then pass the history along if a human is needed.

One demonstration makes the mechanism particularly clear: when waits run high, the caller receives an SMS inviting them into text-based support. The phone has become a doorway to another channel. The customer can chat with a virtual assistant or reach a live agent. For a company with expensive queues and repeatable questions, that changes the design brief from answering every call to finding a workable route for each request.

Then the detour became the contact center

On September 27, 2021, Kore.ai launched an expanded SmartAssist as an end-to-end contact center as a service. It combined automation, an agent desktop and AgentAssist, which helps live agents with customer history and context. Deployment could complement existing telephony or serve as a standalone contact center. Cloud and on-premises options appeared in the launch announcement.

The distinction mattered. A business could approach SmartAssist as an addition to its existing setup rather than making replacement the first decision. Kore.ai positioned it against traditional interactive voice response systems with AI added afterward. That is the vendor’s architectural argument; buyers still have to test whether it improves their own calls.

“giving live agents a single workspace for comprehensive AI-powered assistance.”Raj Koneru · SmartAssist launch, 2021

The agent workspace deserves attention. Kore.ai’s product materials describe omnichannel conversations, peer and supervisor collaboration, and operational dashboards. The audience includes the person answering the customer and the supervisor watching the queue. A successful transfer depends on what those people can see and do after the bot stops talking.

SmartAssist promotional graphic showing a headset-wearing agent and a message about reaching a live human agent
The human is still in the picture. Kore.ai’s handoff marketing puts a headset beside the automation pitch.

The flight leaves before the office opens

Big Holiday, a Japanese travel company, supplies a concrete example. Its May 2024 announcement described staffing pressure, slow responses and uneven service. SmartAssist enabled round-the-clock answering, including questions before early departures and cancellation requests the previous evening.

After four months, the company reported nearly 90% of incoming calls receiving automated responses and a 75% reduction in transfers to human lines. It also identified long questions and insufficient information as recognition challenges it wanted future generative AI capabilities to address.

75%
fewer transfers to human lines

Reported by Big Holiday after four months. A customer result, not a forecast for every deployment.

Here is the distinction worth keeping: answering a call, containing it and resolving the customer’s problem are different events. A buyer should measure all three. Otherwise, a shorter agent queue could conceal customers leaving without an answer. The percentage matters most when paired with evidence that the task got done.

For an operations team, a useful trial would follow one request from beginning to end: the first question, any clarification, the system action and the eventual handoff. Review the conversations that stall as carefully as the ones that finish. Ask whether the agent receives enough history to continue immediately. That exercise makes the trade-off visible. Automation can remove repetitive work, but a transfer that discards information simply moves the work to somebody else. The customer experiences the whole journey, regardless of which department owns each step.

The money belongs to the maker

SmartAssist’s history sits inside Kore.ai. Raj Koneru founded the maker in 2014. Two days after the 2021 SmartAssist launch, Kore.ai announced a $50 million Series C led by Vistara Growth and PNC, plus a separate $20 million credit facility. Those were company financing, not the cost of deploying SmartAssist.

Official portrait of Raj Koneru, founder and CEO of Kore.ai
Raj Koneru, founder of the company behind the queue-routing idea. Portrait: Kore.ai.

The commercial model is enterprise subscription software. Kore.ai’s sales pages invite demonstrations and customized quotations. Partnerships also carry the software into businesses: IIJ Engineering’s 2022 agreement included SmartAssist resale, setup and operational support. For a buyer, the relevant budget includes connecting systems and maintaining service, alongside software access. A demonstration is the beginning of that calculation.

Copy the handoff. Plan the next version.

The transferable lesson is to begin with a repeatable request, give automation a defined job and make escalation easy. Test the receiving agent’s screen. Check whether the customer must repeat information. Compare completion, repeat contact and waiting time before expanding. This is a practical reading of the design, rather than a promise that the same percentages will follow.

The design is less useful when the underlying information cannot support an answer or the request requires judgment that the workflow cannot supply. Human availability still matters. So does maintenance. Kore.ai’s upgrade guide says SmartAssist is nearing end of life and directs accounts toward XO v11, where Contact Center AI joins other modules. It specifies licensing changes and planning for interrupted conversations during an upgrade.

In May 2026, Kore.ai launched its broader Artemis Agent Platform edition. The product family keeps moving. The durable test remains wonderfully ordinary: when someone asks for help, does the next step actually help?