Circles and OpenAI: +22% ARPU and -9% churn in telco

Circles: A timeline of an AI integration built to last

Circles is a digital telco operator founded in Singapore, now present in several Southeast Asian markets. Unlike traditional operators, it built its infrastructure on a cloud-native and data-driven model from the start. This choice made integration with OpenAI technologies smoother than in large legacy telcos.

Over the past few months, Circles has adopted the OpenAI APIs and the Codex model to power an AI-native personalization system. In particular, the objective was twofold: to improve the user experience throughout the entire customer lifecycle and to accelerate the internal development of new features.

Therefore, the project wasn't limited to a standalone chatbot or recommendation engine. Instead, it redesigned how the platform generates, tests, and deploys personalized experiences at scale.

The numbers that define the case: ARPU, churn, and development speed

The results published by OpenAI speak of three distinct metrics. First of all, ARPU — Average Revenue Per User — grew by 22%. This figure indicates that existing customers are spending more, likely because they are receiving offers more relevant to their consumption profile.

Subsequently, the churn rate dropped by 9%. Furthermore, efficiency in the software development process has significantly improved, thanks to the use of Codex for code generation and testing. Therefore, the impact of AI is distributed across multiple business levels simultaneously.

These numbers deserve a critical reading. According to McKinsey Global Institute , companies that systematically integrate AI into core processes achieve up to three times higher results than those using it in a fragmented way. The Circles case confirms this thesis.

Architecture of the approach: What Circles really did

Circles' operating model is based on three main components. First, a unified data layer that collects real-time behavioral signals — data usage, contact frequency, purchase history. Then, a personalization engine powered by OpenAI APIs that generates dynamic offers, messages, and recommendations for each user segment.

Finally, Codex is used internally by development teams to accelerate code writing and testing. This significantly reduces the time-to-market for new features. Similarly, teams can iterate faster on personalization logic itself.

This approach — data, AI applied to the customer, AI applied to development — represents a replicable structure even outside the telco sector. Therefore, it is useful to consider which elements are transferable to Italian B2B or retail contexts.

Winners, losers, and those who watch from the sidelines

The most obvious winner is Circles itself, which consolidates a measurable competitive advantage in a sector — the telco sector — historically characterized by compressed margins and high competition. Furthermore, OpenAI strengthens its narrative of AI applied to business with a concrete and documented case study.

However, the real losers are the operators and companies that continue to postpone AI integration. According to Gartner , by 2027, over 60% of companies that haven't adopted generative AI in customer-facing processes will lose market share to more advanced competitors. The gap widens every quarter.

Those who stand by and watch — and there are still many Italian SMEs — risk finding themselves in a difficult recovery position. Despite this, the good news is that the tools are accessible and the integration models are now documented.

SHM Studio Reading: What it means for Italian marketing

The Circles case is relevant for Italian marketing managers for at least three reasons. Firstly, it demonstrates that generative AI produces measurable results on real business KPIs — not just on superficial engagement metrics. Secondly, it shows that integration must be systemic, not episodic.

Thirdly — and perhaps this is the least discussed aspect — it highlights how AI can act simultaneously on the customer front and the internal front (development, operations). This dual impact is what generates the greatest return on investment.

In SHM Studio we work with B2B and retail companies that want to apply these principles to their own context. For example, AI personalization of content is one of the pillars of our projects. Digital marketing and SEO copywriting . Similarly, we integrate advanced segmentation logic into campaigns Google Ads and Linkedin .

What the numbers don't say: Enabling conditions

A case study like Circles' tends to highlight the results. However, it's useful to ask what conditions made it possible. First of all, Circles already had a consolidated data-driven culture. Then, it had a flexible cloud infrastructure. Finally, it had internal teams capable of integrating external APIs autonomously.

For many Italian SMEs, these conditions are not yet fully in place. Consequently, the path to similar results requires a preliminary phase of data readiness and process review. Therefore, it's crucial not to just copy the model, but to adapt it to your own digital maturity.

According to Harvard Business Review , the companies getting the best results from AI are those that invest first in the foundations — clean data, clear processes, trained teams — and then in applications. Circles is a consistent example of this.

Next moves: Three operational directions for 2026-2027

In light of this case, it's possible to identify some concrete directions for Italian marketing and digital managers. First and foremost, it's worth conducting an audit of your customer data base: data quality, completeness, and accessibility are the prerequisites for any AI project applied to personalization.

Furthermore, it is appropriate to explore the integration of AI APIs into your marketing automation and CRM tools. Many platforms already widespread in Italy — from HubSpot to Salesforce — offer native connectors with OpenAI and equivalent models. Therefore, the starting point is not necessarily custom development.

Finally, it is strategically relevant to start measuring the impact of AI on real business KPIs — ARPU, churn, conversion rate — and not just on operational metrics. This change in perspective is what transforms a pilot project into a scalable investment. I SHM Studio AI services are designed exactly in this direction: from strategy to measurable implementation.

For those who want to delve deeper into how to structure this path within their own company context, the team at SHM Studio is available for dedicated consulting. You can contact us through the contact page or explore our Blog for further insights on AI, SEO, and digital marketing. Among the services available, we also highlight activities of SEO and web development performance-oriented.