We automate 80% of routine FTMO customer requests

What they say about us

,

Challenge

Solution

Collaboration results

FTMO is a global prop-trading platform that processes thousands of messages from traders around the world every day. In an environment of high market volatility and spikes in communication, customer support needs to respond quickly, consistently, and without downtime.

The company now automates 80% of routine communication thanks to SentiSquare’s artificial intelligence, saving several operators time.

Extremely dynamic environment

FTMO is an international fintech platform for forex and cryptocurrency traders, helping talented traders grow.

Political events, regulatory changes or movements in financial markets can cause a spike in client inquiries within a few hours.

Customer support communicates in 21 languages ​​and works with information from several systems, from the trading platform to internal tools to emails and forms.

The rapidly changing situation on the markets brings sudden peaks in communication. On quiet days, an average of about 2000 emails arrive per day, but within the variability of geopolitical events and therefore volatility in the markets, the number can jump up to 5000 messages per day. Every change in the markets is immediately reflected in the behavior of clients and thus in the workload of the team.

Routine inquiries were hampering team capacity

Before implementing SentiSquare’s No-Code NLP platform, FTMO’s customer service team was dealing with several major problems. A large part of the queries were repetitive and followed clear rules. However, operators had to solve them manually, often under time pressure and in different languages. At the same time, it was an environment with high demands on accuracy. Answers depend on specific rules, account status or even regional differences (e.g. USA vs. Europe).

Training new operators was time-consuming and scaling the team was inefficient. Capacity was abundant on quiet days, but lacking during peak times. FTMO was therefore looking for a solution that would:

  • stabilize the team’s workload
  • speed up the response time
  • ensure adequate answers across languages
  • operate fully in an internal, secure environment

A robust system that understands context and language

FTMO chose the No-Code NLP platform SentiSquare built on small specialized language models (SLMs) that run locally on FTMO's own servers.

SentiSquare has created a robust, multi-layered system for FTMO that can understand the context of a query and then respond appropriately. All messages are unified into a single communication pipeline.

The system first understands the context of the query, evaluates its type and complexity, and then:

  • processes routine queries fully automatically
  • prepares more complex cases as a draft for the operator

Responses are always checked in accordance with FTMO's internal rules and metadata (account type, status, history). AI never performs any actions in client systems, it only generates responses according to approved templates.

Today, approximately 80% of routine customer requests are handled automatically. The system responds within minutes, maintains a coherent tone of communication, and handles even peak traffic without operator intervention. It also handles significant fluctuations in communication volume.

SentiSquare automates responses in the five most commonly used languages ​​(English, German, Spanish, French, Portuguese). The solution also includes translation into a single language (English) using local LLM in the No-Code NLP platform, eliminating the need for language specialists for routine queries.

Request detection through multiple connected steps

SentiSquare technology enables automated, secure and accurate customer communication.

When a message arrives from a client, our small specialized language models first detect the type of query or request. This detection includes not only identifying the topic, but also classifying the complexity and potential risk that the message poses.

Based on the detection, a draft response is prepared, which is further checked with metadata and rules stored in the FTMO system, such as the specific rule the client violated, the type of violation or their account history.

The system selects an appropriate template and automatically creates an outbound email, ready to be sent to the client. This makes responses fast, consistent and accurate, while all sensitive information remains internal and the system never performs any actions beyond generating the response.

The final step ensures strict adherence to internal rules, consistency of responses and minimization of the risk of errors.

All data stays inside FTMO

SentiSquare's language models are designed to be as secure as possible for operation in financial and regulated industries.

The entire solution runs fully on-premise. No data leaves the FTMO infrastructure. Each automated response is generated solely based on approved rules and templates. The system never performs any actions in client systems.

This approach ensures 100% process security and complete control over what and when is sent to the customer.

Rapid adaptation to changes and extreme peaks

Thanks to SentiSquare's No-Code NLP platform, FTMO now manages customer support operations in a highly volatile environment without unnecessary stress, team overload or compromises in quality. Automation can save several hundred hours of time per month, thus replacing the work of several busy operators. Customer support today can quickly respond to turbulent market situations that are commonplace in the world of financial markets.

The benefit of the solution is operational stability. FTMO today

  • maintains fast reaction times even under extreme load
  • eliminates routine operator work
  • maintains consistent responses across languages ​​and markets
  • does not have to operationally increase or decrease the team according to market volatility

Do you work in a dynamic industry with changing communications? Are you facing an increase in messages and are you missing people? Or do you have a multilingual team and are you looking for a way to increase efficiency without a linear increase in headcount?

Find out how SentiSquare can automate your routine and give back time to your best people.

Try it out with your own data

More stories
from our customers

Read more