Daily Tech Digest by Kannan Subbiah: Daily Tech Digest

Quote for the day:

"You have to have your heart in the business and the business in your heart." -- An Wang

AI agents can talk — orchestration is what makes them work together

“Agent-to-agent communications is emerging as a really big deal,” ’s chief innovation officer told . “Because if you don't orchestrate it, you get misunderstandings, like people speaking foreign languages to each other. Those misunderstandings reduce the quality of actions and raise the specter of hallucinations, which could be security incidents or data leakage.” ... In another critical evolution in the agentic era, human evaluators will become designers, moving from human-in-the-loop to human-on-the-loop, according to Sanders. That is: They will begin designing agents to automate workflows. Agent builder platforms continue to innovate their , Sanders said, meaning nearly anyone can now stand up an agent using . “This will democratize , and the super skill will be the ability to express a goal, provide context and envision pitfalls, very similar to a good people manager today.” ... Organizations should begin “expeditious programs” to infuse agents across workflows, especially with highly repetitive work that poses bottlenecks. Likely at first, there will be a strong human-in-the-loop element to ensure quality and promote change management. “Serving as an evaluator will strengthen the understanding of how these systems work,” Sanders said, “and eventually enable all of us to operate upstream in agentic workflows instead of downstream.”

Integrating AI-Enhanced Microservices in SAFe 5.0 Framework

can be a game-changer for within SAFe. By optimizing decision analytics and enhancing value stream performance, AI simplifies, rather than complicates. I know what you’re thinking: AI tools can add complexity. One client put this to the test, and we found AI helped reduce the noise. It sliced through the data smog to identify hidden and automate mundane tasks like financial forecasting and risk management. ... Integrating decentralized AI models into SAFe’s can significantly enhance their autonomy. During a high-stakes project, we shifted from a centralized to a decentralized model, which allowed ARTs to self-optimize and adapt to shifting priorities seamlessly. It was like giving ARTs a brain of their own. Decentralized AI models reduce the bottlenecks you'd typically encounter in centralized systems. Think of the ARTs as small startups within the larger enterprise ecosystem, each capable of making swift, informed decisions. ... This isn’t just a tech enthusiast's dream—it's an emerging reality. The maturity of AI technologies spells a future where enterprises aren’t just keeping up; they’re setting the pace. So, if there’s a single, actionable insight to glean from my journey, it’s this: enterprises need to actively pursue cross-industry collaborations, invest in AI-powered microservices, and hone their Agile professionals’ skill sets.

Incorporating Geopolitical Risk Into Your IT Strategy

IT organizations know how to plan for unexpected outages, but even the most rigorously designed strategy is vulnerable to the shifting winds of geopolitics. and technology leaders need to know how their organizations will respond to geopolitical disruptions, and scenario planning needs to be a priority. ... "The IT department can treat geopolitical disruption as an expected operational variable rather than an unforeseen catastrophe. Good and tested enterprise risk management frameworks, investment in government affairs partnerships and ongoing board engagement should start to manage and prepare for this," Dixon said. CIOs need to do around the risks facing their enterprise, and evaluate how IT is teaming with business units, security teams and the CISO on a cohesive tech strategy that builds security, including artificial intelligence security, in from the ground up, said Sean Joyce ... "You're as strong as your weakest link," Joyce said. "As geopolitical risk becomes more prominent, you're going to see tools like cyber being leveraged by countries, particularly those that don't have stronger military or other capabilities. For some, it may be the only tool they can leverage." Physical infrastructure, geography and power supplies are also now areas of risk CIOs need to consider, and infrastructure strategy must align with sustainability, energy realities and geopolitical stability.

Six Architecture Challenges for Startups

The risk is not that the first version is imperfect; that is inevitable. The risk is that the team keeps layering new functionality on top of an accidental architecture. At some point, the cost of change becomes so high that every small modification feels dangerous. The architectural challenge is to intentionally decide where to accept debt and where to invest in structure. Startups need a minimal set of principles – for example, clear domain boundaries, basic , and a simple deployment model – that allow speed without locking the product into a dead end. ... If the product team is still validating pricing models, redefining the customer journey, or experimenting with different verticals, any rigid decomposition can turn into friction. Yet avoiding boundaries altogether leads to a “big ball of mud” that is equally hard to evolve. A practical approach is to use provisional boundaries based on current value streams – onboarding, transaction processing, analytics, etc. – and treat them as hypotheses. The challenge is not to find the perfect structure from day one, but to keep those boundaries explicit and adjustable as the evolves. ... Startups must make conscious decisions about where they are comfortable being tightly coupled to a provider and where they need portability. That requires viewing cloud services through a business lens: What is strategic IP, what is replaceable, and what is pure commodity? Aligning these categories with architectural choices is a non-trivial design challenge, not just a procurement decision.

Platform-as-a-Product: Declarative Infrastructure for Developer Velocity

Without centralized guardrails, teams often compensate by over-allocating resources "to be safe", leading to inconsistent environments and unnecessary cloud spend that is only discovered after deployment. ... What is missing is a developer-friendly abstraction that brings these related concerns together. Developers need a way to express intent (not only what infrastructure is required, but also how the application should be built, deployed, configured across environments, secured, and sized) without having to implement the mechanics of each underlying system. From a platform engineering perspective, this abstraction represents the core of an and can be implemented as a lightweight . ... The platform comprises several interconnected components. coordinate everything, pulling code from repositories, building and unit testing applications (with tests written by developers), checking security, creating cloud infrastructure with , and deploying to clusters with . The configuration controls all of this, telling each component what to do. The architecture clearly separates concerns: the CI pipeline handles code building, testing, and vulnerability scanning. CD pipeline handles deployment: creating cloud resources, updating Kubernetes, and configuring environments.

(Re)introducing Adaptive Business Continuity

is designed to provide a framework that delivers better outcomes when organizations deal with losses. The result may be a reduction in documentation (something I greatly favor) but that is not a stated goal. ... My experience over the years has led me to conclude that trying to define priorities for the resumption of services is wasted effort. Many activities can take place in parallel, and priorities will change when disasters occur. A perfect example is the governmental lockdowns and health authority mandates that followed the emergence of . The result is that demand for products and services changed drastically, upending previous priorities. Priorities may be defined following adaptive principles, but it is not at all a stated component of the Adaptive framework. ... For a number of reasons, I would like to see the word “plan” used a lot less within our profession. Seeing the word “strategy” in its place would be a step in the right direction. Strategy improvement is not, however, a key outcome of Adaptive BC efforts. There is some benefit to having clearly defined recovery strategies, but strategies only provide benefit to competent and empowered teams armed with the resources they need to carry out the mission. For this reason, I always emphasize the importance of focusing efforts on capabilities and consider plans and strategies as little more than supporting tools for any business continuity program. The improvement of strategies and/or plans is simply not an expected outcome of Adaptive BC work.

Exactly What To Automate With AI In 2026 For Faster Business Growth

Most founders automate the wrong things. They start with the flashy stuff, the complicated tools and fancy dashboards, while ignoring the repetitive tasks quietly draining their hours. But you need faster, cleaner growth by removing friction from the activities that actually grow your business. ... You shouldn't embark on a day's worth of admin tasks every time a new client says yes. It will only slow you down. Make it easy for them to pay, get a receipt, complete an onboarding form, and submit the required information. On your end, have the folders, follow-up emails, and team briefings set up without you lifting a finger. Question everything you currently do manually. There is no reason it couldn't be an AI agent handling the sequence. All the tools you pay for already have integrations with each other; You're just not using them. The goal is that you could sign client after client because onboarding takes minutes, not hours. ... AI-generated content is awful when you use it wrong. But that doesn't mean you shouldn't involve AI in your content production process. Content still matters in marketing, whether long-form articles, videos, or social media visuals. You need to be part of the conversation, but only with relevant, authentic material. You cannot outproduce everyone manually, so use automations and retain your human genius for the finishing touches. ... The more your life admin runs on autopilot, the more you free up time and energy for your business.

What is AI fuzzing? And what tools, threats and challenges generative AI brings

The way traditional fuzzing works is you generate a lot of different inputs to an application in an attempt to crash it. Since every application accepts inputs in different ways, that requires a lot of manual setups. Security testers would then run these tests against their companies’ software and systems to see where they might fail. ... Today, generative artificial intelligence has the potential to automate this previously manual process, coming up with more intelligent tests, and allowing more companies to do more testing of their systems. ... But there’s a third angle involved here. What if, instead of trying to break traditional software, the target was an AI-powered system? This creates unique challenges because are not predictable and can respond differently to the same input at different times. ... AI fuzzing can also help speed up the discovery of vulnerabilities, Roy says. “Traditionally, testing was always a function of how many days and weeks you had to test the system, and how many testers you could throw at the testing,” he says. “With AI, we can expand the scale of the testing.” ... Another use of AI in fuzzing is that it takes more than a set of test cases to fully test an application — you also need a mechanism, a harness, to feed the test cases into the app, and in all the nooks and crannies of the application. “If the fuzzing harness does not have good coverage, then you may not uncover vulnerabilities through your fuzzing,” says , staff innovations architect for emerging technologies at .

CISOs flag gaps in third-party risk management

CISOs rank among their highest-impact threats. Vendor relationships touch nearly every core business function, from cloud infrastructure and software development to data processing and AI services. Each added dependency expands the attack surface and increases the number of organizations involved in protecting sensitive systems and data. ... Only a small portion of organizations report visibility across third-, fourth-, and nth-party relationships. Most operate with partial insight limited to direct vendors or a narrow segment of the extended supply chain. CISOs say limited visibility complicates incident response, risk prioritization, and compliance planning. When a breach emerges several layers removed from a known vendor, security teams may struggle to understand exposure, timelines, and downstream impact. ... CISOs report rising regulatory scrutiny tied to third-party cyber risk. Regulatory frameworks place greater expectations on organizations to demonstrate oversight across vendor ecosystems, including indirect relationships. Only a minority of organizations feel ready to meet upcoming requirements without major changes. Most report progress underway, with further work needed to align processes, tooling, and internal coordination. Third-party risk management involves legal, procurement, compliance, and executive leadership alongside security teams. ... At the same time, AI adoption accelerates within vendor risk management itself.

Anti-fragility – what is it and why should it be the goal for your organisation?

That ability to thrive in the face of disruption must become the basis for improved resilience. Modern organisations shouldn’t strive for survival, but for continual improvement. In the , that is crucial. Threat actors are constantly changing tack, targeting new , and executing increasingly complicated . Resilience must therefore move in tandem as an ongoing process of learning and adapting. That is the crux of . It defines systems that thrive and improve from stress, volatility, disorder and shocks, rather than just resisting them. If a security model is only designed to recover, it remains just as vulnerable as before. But an anti-fragile approach actively benefits from each attack, identifying weaknesses, addressing them, and adapting as needed. ... Increasingly, organisations are recognising the value in anti-fragility as a strategy and more will adopt it next year. However, getting there means going beyond regulatory compliance. Compliance lays the foundations from which successful cybersecurity can be built, yet many currently see it as the finished structure. There are several problems with that. Security legislation frequently lags behind the threat landscape, and so the gap between a new threat emerging and a new law coming in to address it can stretch over the course of years. Organisations must therefore understand that compliance doesn’t equal protection.

Daily Tech Digest by Kannan Subbiah

Daily Tech Digest - January 15, 2026