- Orange Business has been using traditional AI for many years
- Now, GenAI is helping the company develop new services while reimagining traditional telecom offerings
- GenAI-driven services can raise top-line growth which has been “flattish” for many years, Orange Business told Fierce
Orange Business Chief Products and Marketing Officer Usman Javaid is undertaking an artificial intelligence (AI) strategy aimed at more than just improving efficiency. In line with industry trends, the company is aiming to use generative AI (GenAI) to create fresh revenue streams.
Orange Business reported +7% year on year revenue growth in Digital, Data and AI services so far in 2024, according to Usman. Backed by new GenAI products launched earlier in the year, GenAI services "are forecasted to drive double-digit annualized revenue growth in the coming years," he told Fierce in an exclusive interview.
The subsidiary of French telecommunications company Orange S.A. has applied traditional AI for years, initially using it for automation applications like network management and fraud detection. Like many telcos, Orange Business has long employed AI-driven smart capital expenditures to predict network capacity needs, and traditional AI has also been useful for identifying fraudulent activities, particularly in voice networks.
But with recent advancements, new GenAI applications are reshaping revenue opportunities in the telecom industry.
"Typically, the telco industry over the many years has been very good at efficiency and productivity gains,” Javaid said. “Our top line is very much flattish, and what we can only do is optimize our bottom line and drive more efficiency out of our businesses. This is good, but this is not the reason I am in this job, right? I’d like to drive more top line growth.”
Raising the top line
The telecom industry has been facing pressure as traditional revenue sources from voice calls and text messages decline, with consumers increasingly turning to internet-based platforms. At the same time, infrastructure costs are rising as companies invest in upgrading networks to 5G technology.
Gen AI is now presenting telcos with a unique opportunity to develop new products and services while reimagining traditional telecom offerings.
McKinsey & Company has said that GenAI adoption could "revitalize profitability for telcos" by enhancing profit margins. Specifically, the global management consulting firm reported that Gen AI could increase earnings before interest, taxes, depreciation and amortisation (EBITDA) margins by 10% by 2029. It also estimated that new GenAI use cases could generate $60 billion to $100 billion in new revenues for the telecom sector.
For Orange Business, this has already manifested in a few ways, Javaid said. Among them, the company’s new GenAI-powered Threat Intelligence Platform enables Orange Business to offer customers a predictive cybersecurity service that can anticipate and mitigate threats.
Its Trusted GenAI Platform, launched in March, supports enterprises that prefer on-premises AI solutions over public cloud. It includes “GPU-as-a-Service” and allows companies to use their own AI models within Orange's infrastructure, which Javaid said appeals to many enterprises in France that are “allergic to public cloud.”
Orange Business is also updating legacy telecom services with GenAI, including giving traditional SMS a makeover. It introduced an SMS-based conversational AI system for a French local council, Javaid noted, where residents can ask questions like “When will my bin be collected?” or report issues, and receive responses generated through AI. The company is also exploring GenAI-powered virtual assistants for voicemail and call handling, allowing businesses to automate interactions and follow-up actions, which can benefit small businesses like plumbing services when they can't answer calls directly.
Orange Business’ GenAI strategy
Orange Business’ GenAI strategy emphasizes partnerships and interoperability rather than building proprietary AI models; a common route for telcos to take. Instead of developing its own large language models (LLMs), the company collaborates with specialized AI providers to integrate diverse models across a “middleware” layer, using OpenAI as one part of a larger ecosystem.
The telco's approach is “highly multimodal and multi-cloud,” Javaid explained. This means it works with multiple AI providers for specific use cases, tailoring each solution to meet unique demands. A prime example is their partnership with LightOn, a French LLM provider, which allows Orange Business to offer LLM-as-a-service for French enterprises that need a fully sovereign solution, hosted on a dedicated GPU cluster within Orange’s private cloud.
In situations like alarm correlation for network monitoring, Orange Business uses off-the-shelf models that are adapted using Orange’s own network data. This strategy allows Orange to draw on a variety of AI models and infrastructures, Javaid said.
For broader use, Orange Business also works with models from companies like Mistral, Meta’s Llama and Google Gemini, incorporating Google Cloud Platform (GCP) resources as needed, he concluded.