Why AI Still Needs Human Judgment

How AI Is Changing the Way Engineering Teams Work

Krishna Sai, CTO at SolarWinds, joins The Tech Trek to talk about one of the biggest shifts happening inside IT and engineering teams: AI is moving people from operators to orchestrators. The conversation goes beyond faster code and automation. Krishna explains why AI is changing how teams think about systems, governance, validation, observability, and the skills technical leaders will need as work moves from manual execution to higher level oversight. Key Takeaways • AI is raising the level of abstraction for IT and engineering teams. The work is shifting from operating systems manually to designing systems that can increasingly run, adapt, and respond on their own. • AI does not automatically reduce workload. In many teams, it changes the type of work by moving effort from execution into validation, judgment, risk management, and governance. • Code generation is only one part of the delivery system. Without testing, security review, observability, and strong engineering process, faster code can create more problems faster. • The best AI outcomes depend on strong foundations. Clean data, connected systems, clear ownership, and resilient architecture matter more as AI becomes part of core workflows. • Technical professionals will need stronger systems thinking, business context, adaptability, and domain understanding as AI changes the shape of day to day work. Timestamped Highlights 00:00 Krishna Sai joins the show and sets the stage for a conversation about AI, IT responsibility, skill gaps, and the latest SolarWinds IT Trends Report. 02:14 Why IT is moving from operator to orchestrator, and what that means for teams that used to spend most of their time responding to tickets and manually managing systems. 04:54 Krishna explains why AI feels different from prior technology shifts. This is not just infrastructure change. It touches individual workflows, jobs, and decision making. 08:56 The messy middle of AI adoption. Teams are getting faster at some tasks, but the workload has not disappeared. It has moved into validation, review, and oversight. 14:46 How AI may force teams to rethink the software delivery cycle, sprint structure, feedback loops, and the speed at which customer issues can be resolved 24:27 Krishna shares how principles from distributed systems, including loose coupling and high cohesion, can help leaders build AI systems that can change without breaking everything around them. Standout Moment “AI is a multiplier. It does not magically fix all your problems. It multiplies your current state.” Pro Tips • Do not measure AI success only by how much faster a team can generate code or complete a task. • Look at the full system around the work, including testing, review, security, observability, and ownership. • Build AI workflows with enough flexibility to swap tools, models, and processes as the technology changes. • Invest in systems thinking and domain knowledge. Those skills become more valuable as execution becomes easier to automate. Call to Action Subscribe to The Tech Trek for more conversations with technology leaders on how AI, data, engineering, and modern systems are changing the way companies build.

Why AI Still Needs Human Judgment

Dan Wald, cofounder and chief AI officer at Sciemo, joins The Tech Trek for a sharp conversation about what AI can and cannot do inside real business workflows. The big question: can AI move beyond quick answers and actually support the messy, context heavy work that still lives in Excel, data teams, and functional expertise? Dan breaks down why consumer style AI has trained people to expect instant answers, why that creates risk inside companies, and why the next wave of AI products needs more than a chat box. It needs context, transparency, guardrails, and humans who understand the work well enough to challenge the output. The conversation also gets into AI agents, coding, entry level talent, narrow workflow specific AI, and why replacing judgment is a much harder problem than replacing repetitive tasks. Key takeaways • AI tools are only useful when they understand the context behind the question, not just the wording of the prompt. • Excel remains powerful because users can see the data, change assumptions, and understand the logic. AI products need to earn that same level of trust. • The best AI workflows are not black boxes. They let users inspect assumptions, challenge outputs, and adjust the answer. • Agents can speed up work, but they still need human judgment, especially when the task requires strategy, constraints, or domain expertise. • AI may change entry level work, but companies still need people who can think critically, solve new problems, and understand why the output is right or wrong. Timestamped highlights 00:40 Dan explains how Sciemo helps consumer brands unify messy data and apply AI to inventory, pricing, assortment, and promotion decisions. 02:30 Why the single prompt experience has changed what people expect from AI, and why that expectation can break down inside the workplace. 04:19 How purpose built AI differs from general AI, especially when the workflow requires context, guardrails, and a clear goal. 07:41 Why Excel is still hard to replace, and what AI systems need to learn from the control and transparency users already expect. 12:57 Dan compares AI agents to unlimited interns, useful for many tasks, but still limited without expert direction. 21:57 The slap chop analogy, and why faster tools do not automatically make someone better at the underlying craft. 31:15 Why predictions about technology and work are so hard to get right, even when productivity clearly improves. A line that stuck “Used properly, they’re great. Used poorly, it’s a very new technology. There will be more mistakes than there are winners.” Practical points worth taking • Do not treat a confident AI answer as a complete answer. • Build AI around real workflows, not generic prompts. • Keep humans close to the assumptions, especially when the decision has business impact. • Use AI to move faster, but make sure someone still understands the logic behind the work. Listen next Follow The Tech Trek for more conversations with founders, operators, and technical leaders building through the next wave of AI, data, and product change.

Why AI Will Not Fix Broken Data Teams

Most data teams do not have an AI problem yet. They have an operating model problem. Mike Doll, VP of Data at Guitar Center, joins The Tech Trek to talk about why analytics teams often become reactive ticket factories, and what it takes to turn data into a true business partnership. As companies push harder into AI, automation, and faster decision making, the foundation matters more than ever. If the data team is buried in scattered requests, unclear priorities, and dashboard maintenance, AI will not magically fix the problem. It may only expose it faster. Mike shares how modern data teams can rethink intake, structure analytics partnerships, separate quick BI needs from deeper analytical work, and create a more consultative model that helps the business answer harder questions. Key Takeaways • AI will not fix a broken data operating model. Teams still need clear intake, trusted data, business context, and a better way to prioritize work. • Data teams become ticket factories when every request is treated the same and stakeholders do not understand what happens after they ask for help. • BI and analytics serve different needs. Quick reporting should be fast and reliable, while deeper analytics requires judgment, framing, and business partnership. • Self service only works when the data foundation is strong. Without that foundation, it can create more confusion instead of more speed. • The future of analytics is not just faster answers. It is better questions, stronger context, and data teams that understand how the business actually operates. Timestamped Highlights 00:41 Mike explains his role leading Guitar Center’s central data organization, including data engineering, analytics, BI, data science, and data strategy. 02:09 How data teams become ticket factories, and why unstructured requests can turn analytics into a black box for the business. 05:29 Why analytics delivery is different from software delivery, and why data teams need closer alignment with business leaders. 07:28 Where self service helps, where it breaks down, and why simple questions need a different model than complex business problems. 09:47 Mike explains the consulting model for analytics teams, with dedicated business partners, stronger dialogue, and shared value creation. 15:35 How AI is changing quick BI workflows, and why harder analytics questions still require human judgment and problem framing. 18:00 How Mike started shifting Guitar Center away from reactive ticket taking by improving intake, visibility, communication, and trust. Line Worth Remembering “The value that analytics teams can bring is answering those hard questions.” Practical Moves For data leaders trying to move beyond reactive analytics, Mike’s advice is to start with the biggest points of friction. That might mean creating a clearer intake process, giving stakeholders visibility into work, assigning dedicated analytics partners to key business areas, or rebuilding trust through fast but meaningful wins. The point is not to add process for the sake of process. The point is to create a data function that can move quickly without losing context, accountability, or connection to business value. Stay Connected Follow The Tech Trek for more conversations with technology leaders on data, AI, engineering, platforms, and the operating models behind modern technical teams.

AI Is Changing Cybersecurity Faster Than Teams Can Keep Up

Cybersecurity is no longer just about keeping attackers out. It is about what happens when they get in. Andrew Rubin, CEO and founder of Illumio, joins The Tech Trek to talk about the speed of modern attacks, why AI changes the security equation, and how companies should think about breach containment, micro segmentation, and guardrails for agentic AI. This conversation gets into a practical shift every technology leader needs to understand. As companies move faster with AI, security teams are being asked to protect more systems, more users, more tools, and eventually more agents. The old idea of perfect prevention is not enough. The better question is how quickly teams can detect, contain, and reduce the impact when something goes wrong. Key Takeaways • Cybersecurity is moving at the speed of technology. As AI accelerates product, engineering, and operations, attackers and defenders are both moving faster. • Prevention alone is not a complete strategy. Andrew makes the case for breach containment, where the goal is to stop a bad event from becoming a catastrophic one. • AI gives both sides more leverage. Attackers can move faster with fewer constraints, while defenders can use AI to automate routine security work and improve response time. • Agentic AI will create a new security challenge. Companies need guardrails that let teams use AI at scale without creating uncontrolled risk. • Cyber budgets need to map to risk. The conversation should start with what risk is being reduced, not what a tool can do. Timestamped Highlights 00:30 Andrew explains what Illumio does and why micro segmentation is really about breach containment. 02:36 Why cyber attacks are accelerating because the rest of the technology world is accelerating too. 04:35 Andrew challenges the idea that any security company can promise perfect protection. 09:46 How agentic AI could help security teams automate mundane work and monitor continuously. 13:28 Why cyber spending often gets misaligned when teams focus on tools instead of risk reduction. 16:55 Where human judgment still matters in cybersecurity, especially during moments of crisis. 20:10 Why large organizations are struggling to let employees use AI aggressively while still putting meaningful guardrails in place. 23:46 The parallel between cloud adoption and AI adoption, and why retrofitting legacy systems is harder than building for AI from the start. A Line That Stuck “Cyber is a math problem. The attackers are going after us, the defenders are trying to prevent it or stop it once it happens, and it becomes a math equation at many levels.” Practical Moves For Tech Leaders • Treat AI as a security and operating model shift, not just another tool rollout. • Start security conversations with risk reduction before product capability. • Look for areas where AI can automate repetitive monitoring and analysis without removing human judgment from high stakes decisions. • Build guardrails early, especially as AI becomes embedded into daily workflows for users and developers. Stay Connected Follow The Tech Trek for more conversations with founders, operators, and technology leaders building the next generation of AI, data, infrastructure, and security systems. Subscribe, follow, and share this episode with someone thinking about how AI changes the way modern technology teams build and protect systems.

Why Data Teams Need Software Engineering Discipline

Kenneth Schwartz, VP of Global Data and Governance at Genmab, joins The Tech Trek to talk about what happens when data teams start applying software engineering discipline to modern data work. As AI raises expectations across the business, the challenge is no longer just building more dashboards or models. It is building data products, governance systems, and engineering cultures that can move from experiment to production in a repeatable way. In this episode, Kenneth shares how data teams can reduce sprawl, create stronger stakeholder alignment, shift governance earlier in the process, and use AI agents to accelerate the data roadmap without simply creating more noise. Key Takeaways • Data sprawl often starts with good intentions. Teams want to move fast, but without alignment they can end up solving the same problem in multiple ways. • Software engineering practices are becoming essential in data. Stable interfaces, data contracts, testing, modular design, and clear ownership help data teams scale with fewer downstream breaks. • Governance works better when it is built into the process early. Kenneth explains why governance should not be treated as a cleanup project after the data already exists. • AI can help data teams move faster, but speed alone is not the goal. The bigger opportunity is using automation to improve quality, reduce manual work, and give teams more time to think. • The future of analytics may depend on better foundations. Catalogs, semantic layers, data marketplaces, and governed metrics can make data more usable across BI, apps, chat interfaces, and agents. Timestamped Highlights 00:00 Kenneth Schwartz joins the show to discuss data engineering, governance, data products, and the growing role of AI in modern data teams. 01:17 Why data is still catching up to software engineering, and how low barriers to entry have created sprawl across dashboards, models, and experiments. 02:55 How stakeholder trust, honest conversations, and change management help reduce duplicated work without slowing the business down. 05:23 The software engineering ideas data teams should borrow, including stable interfaces, data contracts, tests, modularity, and repeatable frameworks. 09:21 Why infrastructure, data, and security teams need a more unified engineering culture as AI and data use cases become more complex. 14:43 What it means to shift governance left, and why governance has to become easier for the people expected to follow it. 20:35 How unstructured data, semantic layers, catalogs, metrics layers, and data marketplaces could change how analytics gets delivered. 24:38 Why faster delivery should not automatically mean more dashboards, more models, or more work products. Standout Line “More is not always better.” Pro Tips • Do not treat every new data request as a net new build. Look for overlap, reuse, and shared definitions before creating another dashboard or model. • Build trust before trying to reduce sprawl. People are more willing to standardize when they believe the data team is helping them win, not just saying no. • Move governance earlier in the lifecycle. Capture ownership, quality expectations, access needs, and context when data is ingested, not months later. • Use AI to accelerate the hard parts of the roadmap, but keep the focus on better decisions, not just faster output. Call to Action Subscribe to The Tech Trek for more conversations with technology leaders building the data, AI, and platform foundations behind modern companies. Follow Amir Bormand on LinkedIn for more clips, takeaways, and episode updates.