The Race Toward Autonomous Networks Begins with Better Data

Worldwide telecom companies have made extensive investments in AI in order to automate their network management operations. However, even though a great deal of focus in the industry has been on the development of AI-driven agents capable of making decisions regarding operations, the Spanish telecom technology company, Kenmei, argues that the true challenge lies somewhere else. As per Kenmei, autonomous networks will only be possible when the operator has a stable data foundation in place.

With its headquarters in Valencia, Spain, Kenmeihas been working on AI-driven products and solutions that will help telecom operators evolve towards fully automated networks. In other words, the company's goal is to facilitate the transition of the industry from a rule-based model to Level 4 network autonomy described by the TM Forum.

New Platform Built on Microsoft Fabric

During the DTW event held in Copenhagen by TM Forum, Kenmei presented their new Network Performance Data Product which runs on Microsoft Fabric and can be accessed via Microsoft Marketplace. This platform allows structuring big amounts of data collected from telecom networks in an organized format.

Previously this year at the Mobile World Congress event, the company showed off their solution based on an intent-driven agentic AI system. In such a system, there is an orchestration agent that understands the request of the operator, chooses the AI specialist needed (troubleshooting, anomaly detection, root cause analysis, or interference management), and forms the strategy of actions. However, engineers are responsible for the approval of each decision before making any changes in the network.

Turning Complex Network Data into Useful Insights

Founders of Kenmei have created another venture known as Ingenia Telecom. This is a venture where they have specialized in the analysis of cell trace data, which later got acquired by Astellia. With this kind of experience, Kenmei has been able to focus on converting large amounts of radio access network (RAN) data into organized data for searching and analysis.

Some of the projects carried out by Kenmei include work done for telecom operators. In Telefónica, Kenmei combined multiple RAN data sources using Microsoft Azure and Databricks. In the case of Swisscom, Kenmei managed to combine information from 21 data sources into one single platform. This has become increasingly necessary for the growth of AI-based operations.

Reducing Deployment Time for Operators

Simplification of the onboarding process of telecom operators is one of the main purposes of Kenmei. Normally, it takes several months to prepare the network data for AI solutions because of system integration, data processing, and security clearance.

The new data product offered by the company will decrease this period from several months to just a few days thanks to ready-to-use, correlated data with ontology which makes AI solutions able to see relations in the network. Although the platform has already been released, Kenmei noted that it is still in its initial stage and has not yet been used commercially.

Facing Competition in a Growing Market

Now Kenmei finds itself having to deal with the competition concerning autonomous networking technologies. Apart from Nokia, Ericsson, and Ciena’s Blue Planet, there are some other companies which have developed AI-based solutions for network automation. Despite the fact that Kenmei managed to receive €4 million as a result of its Series A financing in 2024, and also received investments from Telefónica, MasOrange, and Andorra Telecom, Kenmei lags behind others.

In the future, Kenmei will be concentrating on improving its data platform not only for RANs but also core and transport networks, as well as maintaining its collaboration with Microsoft. According to the company, as telecoms move towards using AI in their operations, success in this endeavor will depend not on the technology itself, but on network data.