Designing innovation we can trust
 

This year, Orange is welcoming even more partners, start-ups, and academic research labs to develop and showcase solutions to some of the biggest challenges facing our digital society. Together, we’re working toward a digital world that’s open, inclusive, and built on trust.

The demos tackle some very practical challenges. There’s AI that can understand a customer’s request and respond on their behalf. Other solutions verify who we’re communicating with to help prevent fraud, while networks can adapt instantly to keep people connected.

Beyond their technical capabilities, these innovations address a strategic priority: helping businesses and individuals adopt technologies they understand, know how to use, and can stay in control of. This is central to our Trust the future strategy.

 


 

 

Giving AI the ability to act, without giving up control
 

AI is no longer simply a tool for answering questions. With what’s known as agentic AI, it can understand a request, reason through it, decide what action to take, and carry it out using a company’s tools.

This opens up new possibilities for customer service. But it also raises an important question: how much autonomy should we give AI?

It’s a question addressed by Agentic Customer Care, a project developed by Orange with Google Cloud and Gemini Enterprise for Customer Experience. The solution can, for example, help a customer find the best option for staying connected when traveling abroad. The AI agent understands the request, identifies the most suitable plan, and can activate it directly if needed. It can then guide the customer through setting up their phone. Unlike a conventional chatbot that follows predefined scenarios, the agent remembers the context of the conversation and adapts its actions accordingly.

By handling routine requests, it reduces waiting times and frees up customer service advisors to deal with more complex situations.

But there are limits to what the agent can do. It operates within a defined scope, following rules and objectives set by the company. Human expertise remains essential to overseeing how it works and retaining control over decisions. The solution has already been deployed by Orange Spain.


Trusted agentic AI: How Orange is putting its ambitions into practice
 
Learn more


Making it easier to choose the right AI model
 

Choosing the right AI model is another important part of building trust. With its Responsible AI Radar research project, Orange is developing a tool to help businesses compare AI models based on how they plan to use them and what matters most to them.

Technical performance is just one consideration. Security and compliance also matter, as do data sovereignty, cost, environmental impact, governance, and ethics. Just as a race car isn’t necessarily the best choice for everyday driving, the highest-performing AI model isn’t always the best fit for every task.

The tool helps businesses identify the best trade-off for a specific business need and understand the reasoning behind that choice. The goal is to help companies deploy AI with a clear understanding of the technologies they’re choosing.

Communications you can trust
 

Trust also means knowing who you’re communicating with. You’ve probably heard about phone number spoofing, scams involving fake bank representatives, and even voice cloning. As fraud techniques become more sophisticated, it’s increasingly difficult to know who’s really on the other end of a call. 
 

 

The Trusted Identity Communication research project is exploring a solution. Its prototype aims to create a trusted space within the calling interface, using a certified digital identity to verify who’s calling. Depending on the situation, verified information such as an address or age could also be shared in a controlled way.

The technology could help verify a customer’s identity when they contact a call center before carrying out a sensitive transaction. It could also make interactions between individuals more secure, for example when arranging to rent a property. The prototype uses IP telephony standards and digital identity wallets such as France Identité. The challenge is to make the technology as easy to understand and as simple to use as possible.


Innovation rooted in academic research

The Trusted Identity Communication project, which aims to verify people’s identities during phone calls, began with the Digital Lab at CentraleSupélec. The research was then taken further through a PhD project with ENSICAEN and the GREYC laboratory. The work focuses on securing the identity associated with phone calls using IP telephony standards and digital identity wallets. Four patent applications have been filed as part of this research.
 


Networks that know how to adapt
 

Trust in digital communications starts with securing the infrastructure behind them. A reliable connection depends on more than how well an antenna or network is performing at any given moment. Networks also need to keep services running when something goes wrong and combine different technologies while anticipating how they’ll interact.

Three demos illustrate this approach at different scales, from private networks and mobile networks to satellite connectivity.
 

1. Private Connectivity Sphere: 

We’re developing interconnected private 5G networks that can be managed remotely. What makes them different is their ability to combine fiber, satellite, and mobile connectivity. An AI agent analyzes network conditions and directs traffic along the most suitable route, while network segmentation and security mechanisms protect communications. Currently being rolled out, the solution could be used for applications such as broadcasting sporting events or providing secure, mission-critical communications for defense operations. 
 

2. Coverage Resilience: 

On mobile networks, Coverage Resilience aims to minimize disruption caused by outages or maintenance. When a site becomes unavailable, the solution identifies the affected area and determines how neighboring antennas can best compensate for the loss of coverage. Before making any adjustments, it uses digital twins to simulate the proposed changes and assess their impact on coverage. The technology is currently being rolled out. 
 

3. Satellite Impact Simulator:

Satellite Impact Simulator explores how terrestrial and satellite networks can work better together. Developed as part of a research project and currently being rolled out, the simulator compares different low Earth orbit satellite constellations. It assesses what they can offer based on geographic areas, use cases, available frequencies, and existing networks. It helps determine which areas could be covered, how consistently services would be available over time, and how satellite connectivity might affect connection quality. It also helps anticipate potential interference with mobile networks, for example when neighboring countries use some of the same frequencies.

 

Building collective trust
 

From AI to networks, developing technologies people can trust means considering trust at every stage, from initial design to deployment.

  • Giving AI the ability to act means clearly defining what it’s allowed to do.
     
  • Making communications more secure requires a way to verify people’s identities.
     
  • Keeping networks running means anticipating incidents and understanding their potential impact.  

Turning a technological breakthrough into a practical solution also takes people with different areas of expertise. At Orange, we bring together our research and innovation teams, specialists in networks and customer service, and technology and academic partners. This year, no fewer than seven major sponsoring partners and several start-ups are taking part in Orange OpenTech.

This collaborative approach reflects the ambition behind our Trust the future strategy. We want to go beyond exploring what technology makes possible by bringing together the expertise needed to develop useful solutions that can be deployed and kept under control.


Key takeaways
 
  • AI with built-in controls: Orange teams are developing AI agents that can handle certain customer requests from start to finish, within clearly defined limits.
     
  • More reliable communications: A certified digital identity prototype aims to combat fraud and make communications more secure.
     
  • More resilient networks that work together: Anticipating incidents, adapting connections, and simulating the potential of satellite connectivity can help keep services running. c
     
  • Open innovation: Orange OpenTech 2026 brings together seven technology partners to accelerate joint innovation and develop solutions that make a difference in France and internationally.