MMOOZZGG.com

AI assistant

In short

A large Russian state-controlled telecom operator said on 30 March 2026 that its AI tools for customer service can automate up to 50% of support requests. The system handles up to 300,000 requests a day in text channels. The article notes that telecom support rests on repeated questions, and the tools were built over four years.

Key ideas

  • A large state-controlled Russian telecom operator says its AI tools can automate up to 50% of support requests.
  • According to the company, up to 300,000 requests a day in text support channels were confirmed in practice.
  • The company introduced two solutions, and the chatbot customer support automation was built over four years.
  • Telecom support questions repeat: no internet, how to pay, when the technician comes, how to change the plan.

Prepared by AI from the article

Chatbot customer support automation: a Russian telecom operator closes half of requests alone

NewsTelecom

Easy2 min read

On 30 March 2026 a large Russian state-controlled telecom operator said that its AI tools for customer service can automate up to 50% of support requests. The system handles up to 300,000 requests a day in text channels. The operator’s press release and CNews reported it. This chatbot customer support automation was built over four years.

On this page · 4 sections

What did the operator introduce?

The company introduced two solutions.

  • The first is a chatbot. A chatbot is a program that exchanges messages with a person in a chat window. This one finds the goal of a request, called the intent, and handles requests in different channels without an operator.
  • The second is a knowledge base: selected company documents that the system searches by context. It prepares answers with the help of AI.

Both systems run in the company’s own infrastructure and are built on Russian software.

What are the numbers?

According to the company, up to 300,000 requests a day in text support channels were confirmed in practice. The system has more than 500 request scenarios and over 400 intents trained on the company’s own data. As a result, up to half of all customer requests were automated.

The operator says this cut request handling time, lowered the load on first-line support and sped up the start of new employees. The team developed the solutions for four years.

What does it mean for telecom companies?

Telecom support rests on repeated questions: there is no internet, how to pay, when the technician comes, how to change the plan. At this operator up to half of such requests already close with no person. Subscribers get used to fast answers from large operators, and a regional provider will have to keep this level with a smaller staff.

For a provider with several operators, another point matters. The base of the whole scheme is a knowledge base and the sorting of requests by topic. The operator has more than 500 scenarios, but a small company usually needs a few dozen frequent questions to start.

What should a provider check now?

  1. Export requests for the last month and sort them by topic.
  2. See which share belongs to the five most frequent questions.
  3. Check whether an agent can see the outage status and the subscriber’s balance in your CRM or billing system. Without it, the agent will answer in general words.
  4. Decide which requests always go to a person: connecting companies, claims and technician visits.
Contact us