6,000 Chats a Month, 40% Resolved Without an Agent: How Mandarin Transformed Customer Support

As Mandarin grew, its support infrastructure became more complex. Requests came in through different channels, automation ran in a separate solution, and B2B partners communicated with the team in group Telegram chats. As these processes expanded, scaling customer support without adding to the team's workload became increasingly difficult.

The company needed to bring these processes into one system while keeping customers' and partners' familiar communication channels and reducing the support team's reliance on disconnected tools.

Before moving to Flomni CS, Mandarin used Zendesk to manage customer requests and a separate solution for automation. The systems were integrated, so support depended on multiple tools and reliable communication between them.

Meanwhile, Mandarin's platform continued to evolve, with new financial products, integrations, and workflows for companies across different industries added to its payment services. Communication volumes with customers and B2B partners grew as well.

As a result, the existing infrastructure became harder to manage. If the connection between individual components failed, automated flows could stop working. Requests from different channels had to be routed to agents and support tiers, with all the necessary information passed along.

The goal was not simply to replace Zendesk with another platform. Mandarin needed to make automation easier to manage, reduce its dependence on disconnected tools, and bring automated flows, agent workflows, and request data together in a single platform.

Based on these requirements, Mandarin's team compared several solutions on the market. They evaluated each solution as a whole: how well it could address current support needs and serve as a foundation for future development.

The team wanted a solution that would combine request handling, automated flows, and analytics in one system while allowing integration with internal services.

After comparing options and testing the platform's capabilities, Mandarin chose Flomni CS. The solution included:

One of the key reasons for choosing Flomni CS was the ability to bring agents, the bot, and analytics together on one platform.

The next step was to migrate existing support processes to the new system and gradually adapt it to the team's actual workflows.

Flomni CS was implemented in stages. The teams reviewed existing support processes, moved them into the new system, and refined individual flows during the rollout.

First, the main support channels were connected to Flomni CS and brought together in Dialogues. Agents began handling requests from the web widget, email, and Telegram bots in a single interface.

Requests from different sources are identified using tags and routed to support groups according to configured rules. Agents no longer need to switch between separate systems: conversations, routing, and transfers between support tiers are all managed within one platform.

Today, around 6,000 conversations a month go through Flomni CS. Most of this volume is distributed across three channels:

The range of channels continues to expand. After the main implementation phase, Mandarin also connected the MAX messenger to Flomni CS, adding another touchpoint for users.

For Mandarin, omnichannel support thus became more than the ability to receive messages from several sources. User requests and B2B partner communication were combined in one process, giving agents a shared workspace for different types of requests without switching between systems.

Next, the team deployed a rule-based bot and Megamozg across all main channels, including the web widget and messaging apps.

All requests pass through the bot. It automatically identifies the request topic, then directs the user to the appropriate flow or transfers the conversation to an agent.

A notable example is handling card-linking errors in the authenticated area of the website.

The process works as follows:

This approach automates repetitive handling steps while keeping agents involved in cases that require further investigation.

The bot does more than serve as the first line of support for common questions. It handles steps that can be standardized and, in other cases, helps prepare requests for agents while preserving the necessary context.

The team also created a dedicated workflow for B2B partners, an important part of Mandarin's customer support communications.

Partners already communicated with the team in group Telegram chats, so Mandarin kept the familiar channel and integrated the bot directly into the existing groups.

Identification uses a client ID and other parameters. Through integration with the company's systems, the following company information is then added to the request:

Some internal communications were automated as well. Request status updates are sent to partners automatically based on internal triggers, so they no longer need to contact support separately each time they want an update.

As a result, Mandarin kept B2B partners' familiar communication channel while incorporating it into the overall support process. Partners do not need to switch to a separate system, and agents receive the context they need to start resolving issues faster.

Mandarin brought agent workflows, automation, and analytics together on one platform while keeping familiar communication channels for customers and partners.

Today, around 6,000 conversations a month go through Flomni CS, 40% are resolved without an agent, and partner Telegram chats have become part of a unified support process.

For users, this means being able to reach support through a familiar channel. For the support team, it means one process for handling, routing, and transferring requests. Repetitive flows are automated, while cases requiring a specialist are passed to an agent with the necessary context.