3‐2.Use Cases of Generative AI in ContactCenters (Call Centers)

As shown above, transcosmos is advancing the use of generative AI with
the goal of shaping the “next‑generation contact center.”To illustrate how these
efforts are progressing today, we will introduce three use cases below

Case 1: AI Generated Conversation Summaries
Case 2: AI Assisted Escalation Handling
Case 3: AI Driven Role Playing Simulations

CX Service Diagram

The first example involves AI converting the conversation between the operator and the customer into text using speech recognition, and then generating a summary of the content.

Because the logs are automatically recorded, the VOC (Voice of the Customer) can be extracted from them, and supervisors can provide feedback to operators based on their interaction details.

In addition, we are also working on automatically generating FAQs using these logs.

 We are currently conducting trials with several clients and the results have been highly positive: performance has improved by an average of 112%, and the time spent on after‑call logging has been reduced by 30–40%.

CX Service Diagram

The second example involves AI handling escalation requests from operators in place of a supervisor.

When an operator submits a question, the generative AI refers to existing Q&A materials, manuals, and other knowledge sources to provide an appropriate response.

This use case is also currently being tested with around ten clients. In the most successful cases, the workload associated with escalation handling has been reduced by as much as 60%.

In addition, through targeted training, the AI’s answer
accuracy exceeded 90% within a short period of time.

As an additional note, the “94% accuracy” shown in the diagram reflects the results obtained when testing only with text‑based Q&A data.

When using document‑based sources, the accuracy is somewhat lower—currently around 70–80% based on our observations.

We expect to further improve the accuracy for document‑based responses going forward.

CX Service Diagram

The third example involves AI conducting role‑playingtr taining with operators in place of a supervisor.

As shown in the diagram above, the operator and the AI perform a role‑playing
scenario, after which the AI evaluates the interaction by identifying the strengths, weaknesses, and any mistakes in the operator’s responses.

The goal of this initiative is to reduce turnover among new operators and improve productivity and quality. However, to be transparent, we are still in the early stages, and concrete results will be accumulated moving forward.

That said, the feedback from both supervisors and new operators at sites where this has already been implemented has been very positive. With that encouraging response, we are moving forward with the expectation that the initiative will lead to strong results.

3-3. System Architecture in Practice

CX Service Diagram

So, what kind of system architecture is actually used toimplement the mechanisms introduced so far?

The diagram above provides a high‑level overview.
Using AWS Cloud, transcosmos built this system in approximately two months.

For example, for telephony systems, there is Amazon Connect; for speech recognition, Amazon Transcribe;for search services, Amazon Kendra; and for generative AI capabilities, Amazon Bedrock.

At this stage, it is still difficult to rely entirely on AI, so we are in a phase where human expertise and AI work together. However,in the near future, the areas that can be entrusted to AI will continue to
expand.
With that in mind, we have high expectations for the future evolution of AWS.

3-4.Future Initiatives by transcosmos Regarding Generative AI

CX Service Diagram

Finally, we would like to share transcosmos outlook on the use of generative AI in contact centers (call centers).

We offer a range of service solutions designed to support the customer journey end‑to‑end, including social operations, digital promotion, website development and management, chat operations, and digital
contact center services. 

However, there are still many sources of “inconvenience” across these front‑end operations. We aim to eliminate these challenges through a seamless integration of generative AI and human expertise.

Customer experience is increasingly shifting toward a world where everything—from purchasing a product to receiving customer support—can be completed seamlessly in one place.

We hope to fully support this evolving landscape and contribute to the business
growth of each of our client companies.

4.Unlock the Power of Generative AI in Your Contact Centerwith transcosmos

By this point, many of you may be thinking that you would like to explore introducing generative AI into your own contact center (call center) as well.

To conclude, we would like to introduce the services offered by transcosmos.

Transcosmos provides “trans‑DX for Support,” a global DX platform designed to optimize customer experience (CX) across all customer touchpoints. We have recently added “trans‑AI Chat,” a generative‑AI‑powered chatbot, to this platform.

With the hybrid chat functionality of “trans‑AI Chat,” customers can receive natural, human‑like responses to their inquiries at any time. In addition, by seamlessly integrating chatbot responses with live‑agent support, the solution enables a high‑quality customer experience.

 Traditional chatbots often required users to select multiple scenarios before reaching a solution,and the extra steps involved frequently led to users abandoning the process midway.

In contrast, with “trans‑AI Chat,” users can simply ask questions in natural language and immediately receive an appropriate response—eliminating the need for complex scenario selections.

In addition, for inquiries that require referencing multiple knowledge sources—such as FAQs, manuals, and other internal documentation—“trans‑AI Chat” can summarize and provide answers across these materials collectively. This helps reduce the time it takes for users to find solutions and contributes to higher customer satisfaction.

Operations are also simpler compared to traditional chatbots, as you only need to update the operational knowledge for generative AI to ensure that responses are based on the latest information.

It also enables seamless integration with live‑agent chat, allowing you to deliver a
consistent, high‑quality support experience to users.

 

Key Takeaways

How did you find it? We hope you now have a clearerunderstanding of the current state and future potential of generative AI incontact centers (call centers). Let's recap the key takeaways from thisarticle.

  • ◎ Japan's largest AWS learning event, "AWS Summit
    Japan,"     was held in June 2024.

  • ◎ ｔranscosmos
    also took part as a presenter, delivering a session titled "The New
    Normal Realized by ｔranscosmos: Improving
    Contact Center Efficiency and Enhancing CX Through Generative AI."

  • ◎ ｔranscosmos
    is actively pursuing initiatives aimed at simultaneously improving customer
    satisfaction and reducing costs     through the use of generative AI.

  • ◎ ｔranscosmosis actively pursuing initiatives aimed at simultaneously improving customersatisfaction and reducing costs through the use of generative AI.

    Case 1) AI summarizes conversations
    Case 2) AI handles escalation responses
    Case 3) AI conducts role-play training

  • ◎ Our goal is to eliminate the various "friction
    points"     in contact center front-line operations through the seamless
    integration of generative AI and human expertise.

We hope this inspires you to leverage generative AI totransform and improve operations at your own contact center.

Contact Us