Daily Tech Digest - May 26, 2025


Quote for the day:

“Don't blow off another's candle for it won't make yours shine brighter.” -- Jaachynma N.E. Agu



Beyond single-model AI: How architectural design drives reliable multi-agent orchestration

It’s no longer just about building a single, super-smart model. The real power, and the exciting frontier, lies in getting multiple specialized AI agents to work together. Think of them as a team of expert colleagues, each with their own skills — one analyzes data, another interacts with customers, a third manages logistics, and so on. Getting this team to collaborate seamlessly, as envisioned by various industry discussions and enabled by modern platforms, is where the magic happens. But let’s be real: Coordinating a bunch of independent, sometimes quirky, AI agents is hard. It’s not just building cool individual agents; it’s the messy middle bit — the orchestration — that can make or break the system. When you have agents that are relying on each other, acting asynchronously and potentially failing independently, you’re not just building software; you’re conducting a complex orchestra. This is where solid architectural blueprints come in. We need patterns designed for reliability and scale right from the start. ... For agents to collaborate effectively, they often need a shared view of the world, or at least the parts relevant to their task. This could be the current status of a customer order, a shared knowledge base of product information or the collective progress towards a goal. Keeping this “collective brain” consistent and accessible across distributed agents is tough. 


Unstructured Data Management Tips

"Unlike traditional databases, which define the schema -- the data's structure -- before it's stored, schema-on-read defers this process until the data is actually read or queried," says Kamal Hathi, senior vice president and general manager of machine-generated data monitoring and analysis software firm at Splunk, a Cisco company. This approach is particularly effective for unstructured and semi-structured data, where the schema is not predefined or rigid, Hathi says. "Traditional databases require a predefined schema, which makes working with unstructured data challenging and less flexible." ... Manage unstructured data by integrating it with structured data in a cloud environment using metadata tagging and AI-driven classifications, suggests Cam Ogden, a senior vice president at data integrity firm Precisely. "Traditionally, structured data -- like customer databases or financial records -- reside in well-organized systems such as relational databases or data warehouses," he says. However, to fully leverage all of their data, organizations need to break down the silos that separate structured data from other forms of data, including unstructured data such as text, images, or log files. This is where the cloud comes into play. Integrating structured and unstructured data in the cloud allows for more comprehensive analytics, enabling organizations to extract deeper insights from previously siloed information, Ogden says. 


Why IT Certifications Are Now the Hottest Currency in Tech

The reasons are manifold. Inflation has eroded buying power, traditional merit-based raises have declined, bonuses are scarcer and 2024 saw a sharp uptick in layoffs - particularly targeting middle management and older professionals. Unlike the "Great Resignation" of 2021, professionals today are staying put - not from loyalty but from caution, and upskilling is the key to ensure their longevity. Faced with a precarious job market and declining benefits, many IT employees are opting for stability and doubling down on internal mobility. According to the Pearson VUE's 2025 Value of IT Certification Candidate Report, more than 80% of the respondents who hold at least one certification said it enhanced their ability to innovate and 70% said they experienced greater autonomy at workplace. Even in regions where pay bumps are smaller, the career mobility afforded by certifications is prevalent. In India, for instance, CloudThat's IT Certifications and Salary Survey found that Microsoft-certified professionals earn an average entry salary of $10,900, with 60% of certified workers reporting pay hikes. "The increased value in certifications underscore their critical role in equipping professionals with the skills needed to excel and advance in their roles. As the industry continues to grow, certifications are becoming essential to stand out and meet the demand for specialized skills," said Bhavesh Goswami, founder and CEO of CloudThat.


Speed and scalability redefine the future of modern banking

To expedite digitalisation, global policymakers are introducing regulations such as India’s Digital Banking Units (DBUs), the EU’s PSD2/PSD3 directives, and the GCC’s open finance guidelines. The growth in non-bank financial intermediaries (NBFIs), which has been both more intricate and wider in scope, in the most recent years, obliges the employing of more effective compliance frameworks and the introduction of better risk management strategies. ... Integrating banking directly into non-financial platforms such as e-commerce is on the rise. Based on a report by Grand View Research, the global Banking-as-a-Service (BaaS) market is expected to reach USD 66 billion by 2030. Retailers increasingly partner with banks for instant, personalised offers and payments via identity beacons, enhancing customer experiences through Gen AI-supported interactions. For example, real-time data analytics and machine learning models are now essential for personalised financial services. Reimagined branch visits are becoming an emerging trend, with branches shifting to high-footfall locations like malls. The store-like experience includes personalised offers and decision aids, including immediate approval for flexible loans, made possible by customer identification based on consent.


5 questions to test tech resilience and build a 90-day action plan

The convergence of AI with existing systems has brought technical debt into sharp focus. While AI, and agentic AI in particular, presents transformative opportunities, it also exposes the limitations of legacy systems and architectural decisions made in the past. It’s essential to balance the excitement of AI adoption with the pragmatic need to address underlying technical debt, as we explored in our recent research. ... While AI enthusiasm runs high, successful implementation requires careful focus on use cases that deliver tangible business value. CIOs must lead their organizations in identifying and executing AI initiatives that drive meaningful outcomes. That means defining AI programs with an holistic, end-to-end vision of how they’ll deliver value for your business. And it means taking a platform approach, as opposed to numerous isolated PoCs. ... The traditional boundaries of IT are dissolving. With technology now fundamentally driving business strategy, CIOs must lead the evolution from an IT operating model to a new business technology operating model. Recent data shows organizations that have embraced this transformation achieved 15% higher top-line performance compared to their peers, with potential for this gap to double by next year.


LlamaFirewall: Open-source framework to detect and mitigate AI centric security risks

One particularly concerning area is the use of LLMs in coding applications. “Coding agents that rely on LLM-generated code may inadvertently introduce security vulnerabilities into production systems,” Chennabasappa warned. “Misaligned multi-step reasoning can also cause agents to perform operations that stray far beyond the user’s original intent.” These types of risks are already surfacing in coding copilots and autonomous research agents, she added, and are only likely to grow as agentic systems become more common. Yet while LLMs are being embedded deeper into mission-critical workflows, the surrounding security infrastructure hasn’t kept pace. “Security infrastructure for LLM-based systems is still in its infancy,” Chennabasappa said. “So far, the industry’s focus has been mostly limited to content moderation guardrails meant to prevent chatbots from generating misinformation or abusive content.” That approach, she argued, is far too narrow. It overlooks deeper, more systemic threats like prompt injection, insecure code generation, and abuse of code interpreter capabilities. Even proprietary safety systems that hardcode rules into model inference APIs fall short, according to Chennabasappa, because they lack the transparency, auditability, and flexibility needed to secure increasingly complex AI applications.


Navigating Double and Triple Extortion Tactics

In double extortion attacks, a second layer is added: attackers, having gained access to the system, exfiltrate sensitive and valuable data. This not only deepens the victim’s vulnerability but also increases pressure, as attackers now hold both encrypted files and stolen information, which they can use as leverage for further demands. The threat of double extortion becomes more severe as it combines operational disruption (due to encrypted data and downtime) with the risk of public exposure. Organizations unable to access their data face halted services, financial loss, and reputational damage. ... Triple extortion expands upon traditional and double extortion ransomware tactics by introducing a third layer of pressure. The attack begins with data encryption and exfiltration, similar to the double extortion model—locking the victim out of their data while simultaneously stealing sensitive information. This stolen data gives attackers multiple avenues to exploit the victim, who is left with no control over its fate. The third stage involves third-party extortion. After collecting data from the primary victim, attackers identify and target affiliated parties, such as partners, clients, and stakeholders, whose information was also compromised. 


The 7 unwritten rules of leading through crisis

Your first move shouldn’t be panic-fixing everything in silence, Young says. “You need to let people know what’s going on, including your team, your leadership, and sometimes even your customers.” Keeping everyone in the loop calms nerves and builds trust. Silence makes everything worse, Young warns. ... Confusion is contagious. “Providing clarity about what’s known, what matters, and what you’re aiming for, stabilizes people and systems,” says Leila Rao, a workplace and executive coaching consultant. “It sets the tone for proactivity instead of reactivity.” Simply treating symptoms will make the problem worse, Rao warns. “Misinformation spreads, trust erodes, and well-intentioned responses become counterproductive.” Crisis is complexity on steroids, Rao observes. “When we center people, welcome multiple perspectives, and make space for emergence, we move from crisis management to collective learning.” ... You can’t hide from a crisis, and attempting to do so only compounds the damage, Hasmukh warns. “Clear visibility into what happened allows you to respond effectively and maintain stakeholder trust during challenging times.” Organizations that delay acknowledging issues inevitably face greater scrutiny and damage than those that address situations head-on.


BYOD like it’s 2025

The data is clear that there can be significant gains in productivity attached to BYOD. Samsung estimates that workers using their own devices can gain about an hour of productive worktime per day and Cybersecurity Insiders says that 68% of businesses see some degree of productivity increases. Although the gains are significant, personal devices can also distract workers more than company-owned devices, with personal notifications, social media accounts, news, and games being the major time-sink culprits. This has the potential to be a real issue, as these apps can become addictive and their use compulsive. ... One challenge for BYOD has always been user support and education. With two generations of digital natives now comprsing more than half the workforce, support and education needs have changed. Both millennials and Gen Z have grown up with the internet and mobile devices, which makes them more comfortable making technology decisions and troubleshooting problems than baby boomers and Gen X. This doesn’t mean that they don’t need tech support, but they do tend to need less hand-holding and don’t instinctively reach for the phone to access that support. Thus, there’s an ongoing shift to self-support resources and other, less time-intensive, models with text chat being the most common — be it with a person or a bot.


You have seen the warnings: your next IT outage could be worse

In-band management uses the same data path as production traffic to manage the customer environment, while logically isolating management traffic from production data. Although this approach can be more cost-effective, it introduces certain risks. If a problem occurs with the production network, it can also disrupt management access to the infrastructure, a situation referred to as “fate sharing.” In these cases, the only viable solution may be to send an engineer onsite to diagnose and resolve the issue. This can result in significant costs and delays, potentially impacting the customer’s business operations. Out-of-band management, on the other hand, uses a separate network to provide independent access for managing the infrastructure, completely isolating management traffic from the production network. This separation is crucial during major disruptions like provider outages or security breaches, as it guarantees continuous access to network devices and servers, even if the primary production network is down or compromised. ... A secure connection links this cloud infrastructure to the customer’s on-premises IT setup, usually through a dedicated private network connection, SD-WAN, or an IPSEC VPN. This connection typically terminates at an on-premises router or firewall, safeguarding access to the out-of-band management network. 

No comments:

Post a Comment