AI in Telecom Operations

I'm excited to announce the launch of AI Refinery for Sovereign and Agentic AI, a groundbreaking platform that deepens our partnership with NVIDIA. This first-of-its-kind platform champions data sovereignty and operational resilience through physical AI, paving the way for enhanced competitiveness in the journey toward agentic AI.   As I've mentioned before, I firmly believe that AI presents a unique opportunity for Europe to reinvent its economy, drive productivity, resilience, and competitiveness, and support future growth. I'm incredibly proud to see the momentum our clients are gaining, including Public Power Corporation, Roche, Kion Group, Noli, and Nestlé.   Nestlé, for instance, is launching a new AI-powered in-house service that will generate high-quality product content at scale for eCommerce and digital media channels. This initiative exemplifies the transformative potential of AI in driving business efficiency and innovation. The expansion of our AI Refinery platform is particularly significant for European organizations, enabling them to accelerate the deployment of AI agents while addressing their sovereignty concerns. This is especially crucial for the public sector and critical infrastructure industries, such as energy, telecommunications, and defense.   We continue to support our clients in maintaining control over their critical data and leveraging innovative AI solutions through this expanded AI Refinery platform. More details here: https://lnkd.in/dvekqfB6 #Noli #Nestle #PublicPowerCorporation #KionGroup #Roche #AgenticAI #AI #Accenture

Want a great AI in customer experience case study? Look no further than AT&T. Here's what happened... AT&T just reported that its AI-driven incident-management system prevented 3.1 million unnecessary technician visits and reduced customer downtime by more than 12 million hours over the past year. These numbers are big and grab headlines for sure, but let's break this down into financials, shall we? (PS - my estimates, not anyone else's.) When I researched industry benchmarks to roll a telecom or fiber truck I came to roughly $150 to $300 in narrower field-ops estimates, with broader all-in estimates ranging from about $150 to $1,000 depending on labor, travel, repeat visits, overhead, and opportunity cost. If these numbers are still accurate, plus taking a swag at avoided contact center calls and other operational costs associated with the reported downtime number, then I would estimate savings to be between $850MM - $1.3B annually. This was a capital-grade operating system investment. AT&T started building the end-to-end incident-management system in 2017, launched it for broadband fiber in 2018, expanded it in 2019, added proactive customer notifications in 2021, added generative AI features in 2022, and added AI agents in 2025. The platform also reorganized about 10 petabytes of data, uses technologies including MongoDB, Azure, Databricks, and Snowflake, and includes more than 30 AI models for predicting and diagnosing network issues. So this likely cost $200M - $350M all-in over the life of the program, and perhaps more depending on how AT&T allocates shared platform, cloud, network-ops, and employee costs. I also modeled annual ongoing run cost at $40M to $90M. So this is likely a 10x-plus annual return after the platform is mature. This just might be the model for service AI than starting with the contact center. AT&T is using operational data to prevent the failure, identify the cause and communicate proactively, rather than waiting for a customer to complain and then automating the response. So the takeaway is that the highest-value customer-service AI may sit upstream from service. Companies should be connecting product, network, fulfillment and operational signals to customer context before spending another dollar on conversational automation. The best contact is often the one the company prevents in the firstplace. Now this begs the question, "What's the value prop of Voice AI solutions like Decagon, Sierra, and others when the customer isn't even calling in?" Feel free to challenge my numbers, but I think I've got them right. If you have a sharper pencil, let me know what you think in the comments. Thankfully, I'm an AT&T customer. Sorry, not sorry, Verizon. Can you hear me now? 😂 #customerexperience #ai #voiceai

What if you could listen to every customer interaction—at scale? For years, contact center leaders have struggled with limited visibility. Most QA teams review only 2-5% of calls, leaving critical insights buried in recordings that never see the light of day. AI-powered Conversation Intelligence changes that. Instead of relying on outdated keyword spotting or manually scoring a fraction of interactions, AI can analyze 100% of your customer conversations, extracting call drivers, sentiment trends, and agent performance insights in real time. Imagine what you could do with that level of clarity. Identify trends before they become problems—spot surges in customer complaints and act before they escalate. Coach agents with precision—understand exactly where improvements are needed, without listening to hours of calls. Optimize automation strategies—pinpoint high-volume, repetitive workflows that are ripe for AI-driven automation. When every conversation becomes a source of insight, your contact center stops flying blind and starts making proactive, data-driven decisions. How would that change your CX strategy?

Service providers are turning networks into AI platforms We’re moving beyond the era of pure connectivity. As Masum Mir points out, this shift needs a new architecture at the service provider edge. With the new Cisco Secure AI Grid with NVIDIA, we’re enabling service providers to run AI inferencing across distributed environments. Essentially, we are turning the telecom edge into an "AI factory" - one where security and data sovereignty are built into the foundation. This is about transforming infrastructure into a monetizable asset that powers real-time intelligence. We’re already seeing industry leaders like AT&T, leveraging the platform for public safety at the Discovery District in Dallas, and SoftBank, building its Telco AI Cloud to support robotics and autonomous transport. These projects show just how fast we can move from concept to real-world deployment when we have the right foundation in place. It’s an exciting time to be building the infrastructure that will define the AI economy.

Telcos Will Build Networks for AI, Not Humans By 2030, autonomous AI agents are expected to exceed one billion active instances globally. These agents will interact with APIs, execute transactions, perform inference at the edge, and operate continuously without human intervention. In parallel, AI-generated content is projected to surpass human-generated content in volume across digital platforms. This shift alters the operating assumptions of human mobile networks. Agents require deterministic latency, persistent low-jitter sessions, verifiable identity, and secure orchestration. Their traffic is structured, continuous, and increasingly upstream heavy. Interaction is machine-to-machine, with growing demands for localized compute and context-awareness. Traditional user-based billing models and best-effort routing are not the best for this profile. Telco infrastructure must expose core capabilities as programmable services. Network slicing must prioritize agent-critical traffic. eSIM and IMSI infrastructure must issue and verify agent identity. Edge compute nodes must support real-time model inference. Session orchestration must scale to persistent A2A and B2A traffic patterns. ( Agent-to-X economics). Foundation model providers and AI platforms will require what Telcos uniquely provide: proximity, mobility context, deterministic routing, and national compliance. The network will shift from moving packets to executing intelligence. Operators that adapt will become essential infrastructure in the AI stack.

I just built my first OpenClaw project! I needed to get my hands on it and build an agent with it - and I did - it's called NetClaw — an AI network engineering agent that operates at CCIE-level depth across routing, switching, security, QoS, MPLS, IPv6, multicast, wireless, and more. 30 skills. NetClaw is built entirely on OpenClaw using what I'd call the "tools as skills" architecture. Each skill is a structured knowledge document that teaches the agent how a network engineer thinks — not just what commands to run, but when to run them, what to look for in the output, and what to do next based on what it finds. The agent connects to live network devices through pyATS skills which abstract the MCP server, executes real show commands, parses real output, and makes real engineering decisions.                  I'm writing this post while simultaneously talking to NetClaw in the VibeOps Forum Slack workspace. I asked it to analyze routing tables and interface states from a live device, generate a Draw.io topology diagram, and create an  image of the network state — all through a Slack message. It did all three. No portal. No ticket. No context-switching. Just a conversation with an engineer that never sleeps.                                                                                    Here's what strikes me about this:                                          We've been automating the wrong layer. For years, network automation has focused on pushing configs and collecting data. Template engines. YAML files. CI/CD pipelines for network changes. All valuable. But they automate the execution — not the reasoning. NetClaw automates the reasoning. It doesn't just run show ip ospf neighbor - it knows to check hello/dead timer mismatches, area ID conflicts, MTU issues causing EXSTART stuck states, and passive interface misconfigurations. It follows the same OSI-model troubleshooting methodology that a CCIE would use.                                                                  Skills are composable. The topology discovery skill feeds into the diagram skill. The security audit skill references the compliance skill. The troubleshooting skill pulls from health checks, routing analysis, and log      inspection. This isn't a monolithic runbook — it's a network of knowledge that the agent traverses based on context.                                                              The conversation is the interface. There's something profound about being in a Slack channel, asking a question in plain English, and getting back a security audit with findings categorized as Critical, High, Medium, and Low complete with CVE benchmark references and specific remediation commands. The barrier between "I wonder if..." and  "here's the answer" has collapsed to the speed of thought. https://lnkd.in/eF-xgM8i

What The Telco Industry Taught Me About Client Experience—And Why It Matters Across All Industries In the telco world, we often say: “The product is the experience.” No matter how advanced the technology, customers remember how we made them feel—especially when things go wrong. That principle holds true in all service industries,  where every interaction is an opportunity to create loyalty—or lose it. As someone who has led client experience strategies and loyalty programs at scale in the telco industry, I’m now looking toward other service-related industries with genuine curiosity and conviction. Why? Because I believe that many of the customer experience innovations in telco can elevate customer experiences across the board. Here are 3 key lessons from telecom industry that can have a lasting impact- 1. Proactive, Not Reactive, Service In telecom, we use predictive analytics to identify when a customer might churn—or when a network issue might cause dissatisfaction—before it even happens. Imagine in other service focused industries such as the hotel and leisure sector; detecting a guest’s dissatisfaction before checkout, or even better, before it begins. By leveraging data from prior stays, booking behavior and feedback, hotels can anticipate needs and intervene early. Proactive hospitality builds loyalty in subtle but powerful ways. 2. Digital Doesn’t Replace Human—it Enhances It The telecom industry was one of the first to scale omnichannel service: live chat, apps, SMS, voice, social—all working together to serve the customer where they are. The leisure sector can benefit from a similar strategy: Self-service check-in for speed In-room chatbots for instant service AI-powered concierge tools to personalize experiences But always with a human safety net. Empathy and warmth remain irreplaceable. 3. Closing the Loop Creates Long-Term Loyalty In telecom, we design VoC (Voice of Customer) programs that don’t just collect feedback—they turn it into action. In the leisure sector closing the loop could mean: Acknowledging guest feedback Making visible improvements Following up personally It’s not enough to say “we’re listening.” You must show it. A Final Thought The core of all service focused industries is the same: people helping people. The right team of people with the same set of drivers and values is key to enhance loyalty and customer delight. However, tools, strategies, and structures we use to do that can—and should—evolve. #telecom #service #guestexperience #customerloyalty #leisure #lifestylemarketing #teamwork

𝐓𝐡𝐞 𝐈𝐦𝐩𝐚𝐜𝐭 𝐨𝐟 𝐆𝐞𝐧𝐞𝐫𝐚𝐭𝐢𝐯𝐞 𝐀𝐈 𝐨𝐧 𝐂𝐲𝐛𝐞𝐫𝐬𝐞𝐜𝐮𝐫𝐢𝐭𝐲 & 𝐏𝐫𝐢𝐯𝐚𝐜𝐲 🔹 Positive Impacts: 1. Phishing Detection: Generative AI systems can analyze patterns in communication and detect phishing attempts by identifying suspicious links, emails, or messages. This can help in preventing potential data breaches and unauthorized access to sensitive information. 2. Incident Response: Real-time communication and coordination during security breaches are facilitated by Generative AI technologies. These systems can quickly analyze and respond to potential threats, minimizing the impact of cyber attacks. 3. Security Training: Generative AI can create highly realistic cybersecurity scenarios for training purposes. This allows cybersecurity professionals to simulate and practice responding to various cyber threats and attacks, enhancing their preparedness and effectiveness in real-world situations. 4. Privacy Enhancements: Generative AI enables the creation of synthetic datasets that can be used to train machine learning models without compromising the privacy of real user data. This helps in developing and testing new cybersecurity solutions without exposing sensitive information. 🔹 Negative Impacts: 1. Advanced Phishing: Generative AI-powered phishing attacks can leverage publicly available data to personalize and customize phishing attempts, making them more convincing and difficult to detect using traditional methods. 2. Deepfakes: Generative AI technology can be used to create highly realistic impersonations, leading to the spread of misinformation and deception. This poses significant challenges for verifying the authenticity of digital content and communications. 3. Data Leakage: There is a risk of unintentional exposure of sensitive training data used by Generative AI systems, potentially leading to privacy breaches and exploitation of proprietary information. 4. Automated Cyber Attacks: Generative AI can be utilized to develop and execute sophisticated, automated cyber attacks that adapt to defensive measures and exploit vulnerabilities at a faster pace than traditional methods. 🔹 The Road Ahead: 1-Regulation: Expect increased government oversight and regulation of Generative AI technologies for responsible and ethical use in cybersecurity and privacy protection. 2-Research and Development: Ongoing research into the implications of Generative AI on cybersecurity and privacy, and the development of effective countermeasures. 3-Public Awareness: Educating the public about the capabilities and potential threats of Generative AI in cybersecurity and privacy. 4-Collaboration: Cooperation among Generative AI developers, cybersecurity experts, and policymakers to address emerging challenges and develop proactive strategies. Image credit: Researchgate Disclaimer - This post has only been shared for an educational and knowledge-sharing purpose related to Technologies. #ciso #technology #learinig #cybersecurity

We’ve seen this story before. 2015: “Self-driving cars in 2 years” 2016: “Radiologists obsolete in 5 years” 2024: Reality check. Now fast forward to Telco & Networking: “AI will run networks autonomously.” “Zero-touch operations will eliminate NOCs.” “Self-healing networks will remove human intervention.” Let’s be precise. In telecom environments—especially large-scale DC fabrics, MPLS cores, and multi-vendor ecosystems—complexity doesn’t disappear. It compounds. AI will transform operations—but not by replacing engineers. It will augment decision-making, reduce MTTR, and surface anomalies faster than humans can detect. What actually works today: • AI-assisted fault correlation (noise reduction across millions of events) • Configuration drift detection and compliance enforcement • Predictive capacity and hardware failure insights • Intelligent automation pipelines (closed-loop, but supervised) What doesn’t (yet): • Fully autonomous networks without human guardrails • Black-box AI making production-impacting decisions • One-size-fits-all models across heterogeneous telco stacks The takeaway for Telco leaders: 👉 Don’t chase hype cycles—engineer for reality 👉 Focus on incremental AI adoption with measurable outcomes 👉 Build observability + data pipelines first, AI later 👉 Keep humans in the loop—especially for critical control planes AI in networking is not a replacement strategy. It’s a force multiplier for operational excellence. #Telco #Networking #AI #NetOps #Automation #NOC #Observability

Telecom didn't adopt AI by choice. Networks simply outgrew human capability. For decades, Telecom networks were powered by rules, scripts, and highly skilled NOC engineers. And for a long time, that worked. But today's networks have reached a scale where human expertise alone is no longer enough. A single 5G cell site can generate thousands of performance counters every minute. Multiply that across thousands of sites in a metro network, and you're dealing with millions of events every hour. No operations team , no matter how experienced can analyse, correlate, and respond to that volume in real time. That's where AI stopped being optional. The Evolution of AI in Telecom :- ✅ Rule-Based Automation 🔹 If-this-then-that logic 🔹 Static thresholds 🔹 Automated scripts 🔹 Reliable, but limited to predefined scenarios ✅ Machine Learning 🔹 Learns patterns from historical data 🔹 Detects anomalies 🔹 Predicts traffic demand 🔹 Forecasts network KPIs ✅ Deep Learning 🔹 Understands complex, high-dimensional data 🔹 RF fingerprinting 🔹 Image-based tower inspections 🔹 Sequence modeling for network intelligence ✅ Generative & Agentic AI 🔹 Understands natural language and operational context 🔹 Generates configurations and Root Cause Analysis (RCA) 🔹 Assists engineers with troubleshooting 🔹 Progressively moves toward autonomous decision-making with human oversight 💡 Where Is Telecom Heading? Each wave builds on the previous one. The destination remains the same: ✅ Self-configuring networks ✅ Self-healing networks ✅ Self-optimizing networks The vision is no longer just automation. It's autonomous network operations. AI isn't another feature being added to telecom. It is rapidly becoming the operating system that will power the next generation of intelligent networks. #Telecom #AI #ArtificialIntelligence #AgenticAI #GenerativeAI #5G #NetworkAutomation #DeepLearning #MachineLearning #Telecommunications #DigitalTransformation #FutureOfWork #NetworkEngineering #Innovation