Latest news with #DeepTalks
Time of India
04-07-2025
- Business
- Time of India
The SOCs isn't just a function in the age of AI Era by Dr. Yusuf Hashmi
HighlightsWhy SOC fatigue is a systemic risk, not an analyst issue The role of AI, agentic models, and automation in optimizing MTTR How to design SOCs that scale with relevance, not just volume The intersection of DPDP, data lineage, and SOC accountability The irreplaceable role of human context in an AI-augmented security world In this DeepTalks session, Dr. Yusuf Hashmi, Group, CISO at Jubilant Bhartia Group, reimagines the SOC, tackling AI-assisted triage, alert fatigue, data governance, DPDP liability, and the rising cost of log inflation, to present a bold, practical vision for future-ready security the dimly lit war rooms of cybersecurity, the Security Operations Centers (SOCs), thousands of alerts blink on screens every minute. Analysts scan dashboards, eyes darting, trying to distinguish between noise and the one anomaly that could bring an enterprise to its knees. But in today's AI-fueled world, even these battle-tested security models are showing signs of exhaustion. 'It's time we stop seeing the SOC as just a dashboard of alerts,' says Dr. Yusuf Hashmi , Group CISO at Jubilant Bhartia Group, in a gripping and wide-ranging conversation with ETCIO DeepTalks. 'We must reimagine it as a cockpit, one that is predictive, autonomous, and human-aware.' Dr. Hashmi isn't just describing a shift in tools. He's championing a cultural and architectural transformation, one that demands leadership rethink how security operations are structured, automated, and governed. The breakdown begins The conversation opens with a blunt diagnosis: the traditional SOC is broken. 'There used to be a handful of firewall logs coming in. Today, we're ingesting data from 60-70 different log sources,' Dr. Hashmi explains. 'From endpoints to proxies, from cloud to identity - the ecosystem is sprawling. And each of these sources needs contextual use cases. But most organizations aren't ready for that.' This, Dr. Hashmi says, creates the perfect storm for alert fatigue, a silent killer in cybersecurity. Analysts are overwhelmed, incidents are missed, and trust in the SOC dwindles. AI's promise and pitfalls Dr. Hashmi sees AI not as a silver bullet, but as a powerful enabler, if implemented wisely. 'AI can triage, correlate, enrich. It can suppress false positives and help prioritize what matters. But AI must be trained. It doesn't mature out of the box. You need 5 to 6 months, sometimes longer, to adapt a model to your data,' Dr. Hashmi warns. Dr. Hashmi emphasizes the agentic model, using AI-powered agents to take over repetitive, mundane triage tasks so human analysts can focus on critical decision-making. But the contextual layer, he insists, must remain human. Dr. Hashmi also says 'AI can automate. But it cannot replace the analyst's gut instinct, their ability to think outside the box. That's irreplaceable.' Integration nightmares & log inflation At the heart of SOC dysfunction lies a quietly growing monster: log overload. 'Many organizations don't understand what they're ingesting,' Dr. Hashmi says. 'EPS (Endpoint security) peaks go through the roof. And half those logs? They're noisy. They're being stored, processed, and paid for, but they add no value.' Dr. Hashmi's advice: optimize for relevance, 'You don't need everything. You need what helps you correlate, detect, and respond . Everything else is an expensive distraction.' From alert fatigue to MTTR anxiety Metrics like Mean Time to Detect (MTTD) and Mean Time to Respond (MTTR) have become the new holy grails of SOC performance. But as Dr. Hashmi points out, they're only as good as the underlying architecture and logic. 'If you don't fine-tune your rules, if your alerts aren't contextualized, your MTTA and MTTR suffer. Analysts waste time chasing irrelevant noise, and that one critical alert gets buried.' The fix? Smarter alerting. Better enrichment. Fewer false positives. And yes, more AI-powered correlation engines that understand behavioral baselines. The compliance curveball: DPDP's impact on SOCs With India's Digital Personal Data Protection (DPDP) Act coming into force, Dr. Hashmi sees new pressure on SOC teams especially around personal data ingestion. 'If your SOC is processing DLP logs, you may be dealing with personal data. That means you're accountable under the DPDP. You need governance, visibility, and traceability.' He calls for greater attention to data lineage, understanding where data comes from, how it's stored, who accesses it, and how long it remains within systems. 'Security without governance is a ticking bomb. You need to know your data trail end to end.'notes Dr. Hashmi. SOC design: It's not about tools. It's about context. When asked what makes a modern SOC truly effective, Dr. Hashmi offers a precise and measured answer: Scalability: The platform must handle peak Use Cases: MITRE ATT&CK-aligned rules save Analysts need intuitive, investigation-friendly Awareness: Know your licensing model EPS vs Clarity: MTTR, MTTD, FP rates these are your compass. But Dr. Hashmi's quick to emphasize that no model fits all. 'You must understand your environment. Your threat landscape. Your business impact. No Gartner quadrant can define your context better than you.' The ROI dilemma and the AI hype trap Every CISO today is asked the same thing: What's the ROI on security? Dr. Hashmi believes it starts with asset valuation. 'If you don't know the value of what you're protecting, how will you measure loss? Understand your assets. Quantify their downtime impact. Then map your SOC outcomes against that.' He also cautions against AI - FOMO 'Many CISOs buy AI tools just because they're trending. But if your MTTA isn't improving, your response time hasn't dropped. What did you really gain?' says Dr. Hashmi. On MDRs, cloud SOCs, and cost-efficient architectures For organizations lacking in-house expertise or infrastructure, Dr. Hashmi recommends SOC-as-a-Service or Managed Detection & Response (MDR) models. 'Not everyone needs an on-prem SOC. If you're a smaller firm, MDR can be a life-saver, no licensing, no infra management, no staffing nightmares.' Dr. Hashmi also advocates for cloud-based SOCs with high availability and easy scalability, especially when uptime and redundancy are mission-critical. In perhaps the most poignant part of the conversation, Dr. Hashmi speaks of the unsung heroes of the SOC, the analysts. 'They run 24x7. They're the stars of the security function. But we overload them with Excel reporting, compliance checklists, and fatigue. That has to stop.' pointed Dr. Hashmi. Dr. Hashmi also urges CISOs to sit with their SOC teams, understand their world, and build empathy into governance. 'The SOC isn't just a function. It's your shield. If you love it, you'll nurture it.' In a world increasingly driven by automation, Dr. Hashmi reminds us that passion still powers the best defenses. 'SOCs are like goalkeepers. They don't get applause until something goes wrong. But they're your last line of defense, and your first line of attack.' To modernize a SOC, organizations must combine the power of AI with the wisdom of human intelligence, supported by architecture that scales, data that's governed, and leadership that listens. Because in cybersecurity, it's not just about fighting threats, it's about earning trust, concludes Dr. Hashmi.
Time of India
23-06-2025
- Business
- Time of India
Inside AI Co-Pilots as productivity gamechanger: Rohan Chhatre, Volkswagen Group
AI copilots are transforming enterprise productivity, from coding and customer support to HR, finance, and offer context-aware intelligence, acting as digital colleagues that understand your data, tools, and must come with guardrails: CIOs need to prioritize data readiness, compliance, and change won't replace jobs, but will reshape them, freeing employees from routine tasks and unlocking time for innovation. In today's rapidly evolving enterprise landscape, artificial intelligence is no longer a backroom experiment or a shiny add-on. It is now central to how organizations build, operate, and grow. At the forefront of this shift are AI copilots, context-aware digital assistants designed not just to automate tasks, but to augment human potential. In a compelling DeepTalks conversation, Rohan Chhatre , at Volkswagen Group Technology Solutions India, unpacks what this evolution means for enterprises, leadership, and the future of work. Rethinking work with AI copilots AI copilots are not a futuristic dream, they are already reshaping the way organizations think about productivity and collaboration. Chhatre sets the tone, 'AI copilots are not just intelligent. They are contextually intelligent,' he emphasizes. These digital colleagues are capable of understanding the business context, whether that's an organizational policy, project workflow, or customer interaction, and responding with nuanced, relevant support. Think of an AI copilot as a hyper-organized team member who knows every file location, meeting detail, and deadline. But unlike static automation, copilots adapt, learn, and evolve with usage, making them indispensable partners in both routine operations and complex decision-making. 'AI copilots are like digital colleagues, context-aware, intelligent, and always ready to assist. They know where every file is, what deadlines are coming up, and how to help you work smarter', states Chhatre Inside the enterprise use cases Across industries, the use cases are multiplying. Rohan outlines three core areas where AI copilots are already proving transformative: Software Development: Tools like GitHub Copilot are acting as real-time pair programmers, handling documentation, code reviews, and even refactoring, freeing up developers for more strategic and creative Support: Copilots can unify fragmented data across ticketing systems, emails, and knowledge bases to deliver fast, contextual responses to customers and agents and Business Functions: From summarizing interview notes to drafting employee communications, copilots can streamline administrative tasks that are critical yet time-consuming. Other promising areas include finance (detecting payment anomalies), procurement (pattern recognition), and sales (dashboard generation and trend analysis). 'DevOps and customer support are leading the adoption curve, GitHub Copilot is a game-changer for developer productivity, and AI copilots are transforming how we resolve customer issues', says Chhatre. Redefining the CIO's mandate For CIOs and tech leaders, the AI Copilot era presents a mandate that is both thrilling and daunting. 'If you're still on the fence, you're already behind.' Chhatre offers a three-pronged checklist for CIOs embarking on this journey: Data Readiness: 'Your copilots are only as smart as the data you feed them.' Enterprises must clean, structure, and contextualize their internal datasets to unlock meaningful copiloting & Compliance: With great power comes greater responsibility. Copilots must operate within tightly defined guardrails to ensure privacy, compliance, and enterprise Management: Tech alone won't drive adoption. Organizations must cultivate evangelists, early adopters who showcase the benefits, bust myths, and build cultural trust. 'If you're still on the fence about AI copilots, you're already behind. But adoption must come with guardrails, data readiness, security, compliance, and change management', observes Chhatre Building a responsible AI culture Rohan debunks a common myth: that copilots will replace human jobs. 'Instead of fearing replacement, think delegation,' he urges. Identify the 20-30% of tasks in every role that are repetitive or cognitively draining. Then, let copilots handle them so that humans can focus on creativity, problem-solving, and innovation. 'AI won't take your job. But someone using AI might. It's time to embrace copilots and delegate the mundane, so you can focus on what truly matters,' says Chhatre However, Chhatre also warns against blind trust. 'Copilots hallucinate. They make mistakes. Human oversight is critical.' As such, CIOs must define governance frameworks that include: Ethical model designBias mitigationHuman-in-the-loop supervisionAudit trails and access controls 'Copilots can be wrong, they hallucinate. Don't delegate decisions to AI. Use it to assist, not replace. Keep humans in command,' Chhatre emphasizes. The Human-AI symbiosis Perhaps the most profound insight from the conversation is the vision of the next-gen digital workforce, not one where humans and machines compete, but one where they collaborate. 'That line between human work and copilot work is already thinning,' Chhatre notes. 'Old automation was like pressing a button on a vending machine. AI copilots are different, they understand you're thirsty, know your preferences, and hand you the right drink before you ask.', reflects Chhatre The most successful copilots will be those that embed seamlessly into existing platforms like Microsoft 365, Salesforce, or SAP. Their strength lies not in complexity, but in usability and design. 'User-centered design, emotional tone, multi-language capabilities, these factors will decide which copilots become indispensable,' he adds. 'The copilots that work best are the ones embedded right into Gmail , Outlook, Word, right where you already are. You shouldn't have to go somewhere else to use them', he adds. On the ground, DevOps and customer support teams are already reporting efficiency gains, with early metrics showing up to 25% increase in productivity. The future looks even brighter as copilots move beyond backend tasks to shape sales strategies, legal reviews, and even leadership decision-making. How a Copilot transformed support efficiency At a global enterprise working across automotive and manufacturing verticals, the customer support team faced a familiar challenge, resolving tickets that spanned multiple disconnected systems. Knowledge bases lived in one platform, past ticket logs in another, and product documentation in yet another. Support agents were spending 30-40% of their time manually searching for relevant answers, often leading to delayed responses and inconsistent customer experiences. The company deployed an AI copilot trained specifically on its internal documentation, ticket history, and FAQs. Within weeks, the results were visible. Agents could now simply prompt the copilot with a query and receive a pre-validated, cross-referenced response in seconds. Impact: Average ticket resolution time dropped by 25% Agent satisfaction scores rose, as repetitive hunting for answers was eliminatedEscalations reduced due to more consistent and faster replies Customer support is a great use case for copilots. They can pull from different knowledge bases and give a singular view to the support agent. It helps reduce time and cognitive load. Rohan also mentioned: Old ticket data, knowledge base, workflow logs, copilots bring them together for better resolution. Looking ahead: The roadmap and the realization Over the next 12-24 months, Rohan predicts a surge in enterprise copilots tailored for specific functions, legal, procurement, HR, and sales. These copilots will summarize lengthy reports, surface hidden anomalies, and drive faster insights at scale. But success will hinge on three pillars: Strong strategy with short sprints - Think six months ahead, not three years. Upskilling at scale - From prompt engineering to ethical AI design, literacy must go mainstream. Measurable impact - Usage metrics, productivity benchmarks, and user feedback loops must be built in from day one. Chhatre concludes, 'Lead with purpose. AI copilots are not just a trend, they are a strategy. And strategy needs alignment, clarity, and recalibration.' In the age of copilots, CIOs are no longer just technology stewards. They are architects of intelligent collaboration, custodians of ethical innovation, and enablers of the human-machine partnership.
