Altice Portugal aims for intelligent network operations

José Pedro Nascimento, Chief Technology Officer at Altice Portugal, is leading the telco’s transition to intelligent network operations, or “AINetOps.” Speaking at Mobile Europe’s recent Network Now event, he shared the operator’s position on improving customer service through artificial intelligence (AI) and automation, some of the early use cases for AI-driven operations, and some of the obstacles.

He said that the “most important challenge” — not just for his organization, but for all operators — is process simplification. That is, getting people to stop using the old manual processes that they are used to and adapt to new automated tools. Indeed, “simplification” has become a mantra for Nascimento in the telco’s transition to intelligent operations.

When faced with new automation projects, he said “people tend to keep things as they are in terms of the way of working ... It’s not the right way, but it’s how people tend to act. If we don’t simplify, we can use AI and whatever tools we want … but we will stay the same and we won’t evolve.”

Another challenge is lack of skills, such as in data science and AI. Especially for an incumbent operator like Altice Portugal (formerly, Portugal Telecom), he said, “we need new skills,” and noted that that the telco has started programs to “change the mindset of our people” and “to get them engaged in these new transformation projects that we have.”

Altice Portugal is the country’s largest telecom operator with 6 million mobile subscribers and 1.7 million fixed-line customers (of which 1.4 million are fiber customers). Its 5G network covers 95.6% of the population and it has rolled out fiber to 6.3 million premises. The telco was acquired by telecom investor Patrick Drahi’s Altice in June 2015. It is a turbulent time for the Portuguese operator as it navigates allegations of financial misconduct that emerged in July this year, while Altice is also reportedly considering a sale of the assets to reduce debt.

Customers in focus

Nascimento described AINetOps as a unifying concept that brings together customer experience, network intelligence and “cognitive operations.” It encapsulates an envisioned end state in the transition from legacy, siloed, manual operations for fault and performance management.

The operator wants to deliver “quality of experience” from the user’s perspective, giving customers “transparency on service usage”; enabling online configuration; and using proactive maintenance, for example.

“Everything that we can do for self-healing and self-care is massive,” he said, explaining that the good image customers have of the operator can be “destroyed” when they have a service problem and encounter difficult processes to correct it.

Improving customer experience further, network intelligence and cognitive operations would enable Altice Portugal to use AI-based insights to proactively detect performance issues and automate responses to resolve them in a closed-loop operation.

Building blocks for automation

Many different components are needed for the telco’s autonomous operations vision. For Nascimento, “everything starts with a very good catalog and inventory of our network. This is the basis.”

In addition, there also needs to be “sensing” in place for monitoring the status of networks and services; “intelligent analytics” to correlate the KPI data and perform predictive maintenance using AI/ML decision engines; and “acting” for service configuration and activation.

Together, these capabilities create autonomous operations, characterised by the collection and processing of large amounts of data, automated analysis and actuation, and closed-loop operations.

Putting it to work with use cases

Nascimento shared examples of some of the AI-based solutions and use cases that have been developed by Altice Labs. Previously known as PT Inovação, this was renamed Altice Labs in 2016 and became the research and development centre for the Altice Group. Headquartered in Aveiro, Portugal, it has seven other lab offices in the country as well as two locations in Brazil.

• Smart WIFi: Built by Altice Labs, the operator is exploring a solution that uses AI to monitor the WiFi traffic in customers’ homes and recommend the best locations for placing WiFi extenders to improve coverage throughout the premise.
• Network Pattern Identification: A proactive monitoring use case, it enables Altice to correlate data from various customer premises equipment (CPE) to detect unusual network patterns that can affect customers’ services. Altice has an array of CPE, including its own fibre gateways and legacy set-top boxes (STBs) from multiple suppliers using different operating systems.
  “This is a very interesting use case,” he said. “We have a lot of pieces of hardware and software at home, and this provides a set of information that is brutal. If you can correlate all this information, and we’re doing this, we’re finding new patterns that we couldn’t find before. And then, we can have some proactive maintenance — we can say to the customer, that STB will fail in a few days, or that fibre gateway needs to have a new configuration or reset.”
• Autonomous Recursive Intelligent System (ARISE): An AIOps module developed for Altice Labs’s OSS platform NOSSIS One, it uses advanced machine learning to autonomously identify root causes in a cascade of failures, enabling faster resolution and better network insights. ARISE complements Altice’s fault management analytics with AI.
  “We are getting better results because the correlations are much more thorough and we can be more precise on the action, on what we have to do,” he explained.
• Analytics for cyber security: The operator is also working on use cases for AI and “unsupervised ML” algorithms to detect anomalous user and entity behavior as part of its security measures for perimeter and end-point protection.