Circles нарастил ARPU на 22% и снизил отток на 9% с помощью OpenAI API и Codex

Telecom operator Circles increased average revenue per user (ARPU) by 22% and reduced customer churn by 9% by migrating its services to an AI-native platform built on the OpenAI API and Codex, according to the official OpenAI blog.

What the AI-native platform delivered

Circles builds telecom services as AI-native from the ground up: rather than adding artificial intelligence on top of legacy infrastructure, the company uses OpenAI models as the foundation of its product. A traditional operator adds AI as an additional layer on top of its network and tariff plans — Circles makes the intelligent system the core, with the network as the infrastructure beneath it.

Two OpenAI tools are used to achieve this. The OpenAI API powers personalization: the system analyzes subscriber behavior and offers the right plan or service at the right moment. Codex — OpenAI's automated code-writing tool — accelerates the development team's work by reducing time spent on routine tasks and allowing engineers to focus on more complex challenges.

According to the OpenAI Blog, the results of the implementation were recorded across three metrics:

• ARPU (average revenue per user) grew by 22%
• Customer churn decreased by 9%
• Development efficiency improved: the team ships new features faster

"Circles uses the

OpenAI API and Codex to build AI-native telecom services, increasing ARPU by 22%, reducing churn by 9%, and improving development efficiency," states the official OpenAI blog.

Why these numbers matter for telecom

Telecom operators worldwide face the same problem: voice and mobile data have become a standardized commodity, with competition driven primarily by price. Retaining subscribers requires constant expenditure, and acquiring a new customer costs significantly more than keeping an existing one.

A 9% reduction in churn means that out of every 100 subscribers ready to switch to a competitor, 9 stayed — directly due to AI-powered personalized offers. A 22% increase in ARPU shows the other side: when the algorithm offers the right package at the right time, customers pay more — not under pressure from a salesperson, but because the offer matches a real need.

Accelerating development with Codex creates a third competitive advantage: the operator brings new features to market faster and does not lose pace due to routine boilerplate code. This allows Circles to iterate faster than traditional competitors.

The combined effect of the three changes — revenue growth, churn reduction, and faster development — means Circles is improving all key competitiveness factors simultaneously. According to the OpenAI Blog, the AI-native architecture allows the company to achieve this without compromising between speed and service quality.

What this means

The Circles case study published by OpenAI confirms: the AI-native approach in telecom is not a marketing concept but a strategy with measurable business results. Three key metrics — ARPU +22%, churn −9%, faster development — improved simultaneously through the OpenAI API and Codex. For competitors, this is a signal: falling behind in AI adoption affects not only technological image but also revenue and customer retention.

Frequently asked questions

What OpenAI tools does Circles use?

According to the official OpenAI blog, Circles uses two tools: the OpenAI API for personalizing telecom services, and Codex for accelerating code writing and review within the development team.

What results did Circles achieve after implementing AI?

According to the OpenAI Blog, Circles recorded three measurable indicators: ARPU grew by 22%, customer churn decreased by 9%, and development efficiency improved significantly.