The episodes of the Innovantage podcast hosted by Sigli’s CBDO Max Golikov cover a wide range of topics, including technology, business, and the role of digitalization. In the spotlight of the new episode are not just technologies, but also people. How can business leaders successfully implement innovations without facing resistance from their teams?
Thiago de Faria, Senior Solutions Architect at AWS and a recognized Team Topologies expert, shared his perspective on this and many other important questions.
Thiago has spent the past decade at the intersection of data, distributed systems, and organizational dynamics. Through years of hands-on experience, he came to understand that success isn’t just about algorithms or tools. It’s about communication and how teams work together.
This insight led him to pursue leadership roles, including such positions as CTO, director, and team lead. He later joined AWS, where he led startup solutions architect teams, working closely with early-stage startups. After a period of freelancing, Thiago returned to AWS. Now, he focuses on enterprise modernization. His goal is to help organizations realize that technology is rarely the real challenge in their transition. They need to pay more attention to everything else around it.
For Thiago, building sustainable businesses goes far beyond technology. It is about people and communication. Success depends on understanding human behavior, managing egos, and being kind.
He doesn’t believe in the top-down leadership model and the "do it because I'm the boss" mindset in the modern world. Today’s teams are motivated by more than just money or fear. The traditional stick-and-carrot approach no longer works. Instead, leaders must tap into what truly drives people and apply that insight to how technology and organizations evolve.
Thiago mentioned Team Topologies as a powerful framework for structuring teams and improving how they collaborate. It was developed and described by Matthew Skelton and Manuel Pais. The framework is based on the ideas of DevOps, Agile, Lean, Deming’s principles, and the Theory of Constraints. Its core goal is to enable fast flow from idea to production, while still meeting security, compliance, and feedback needs.
Overplanning or overcommunicating can also be dangerous for team efficiency. Team Topologies introduces a shared language to design team interactions intentionally. Without this structure, teams often drown in context, which can result in cognitive overload and a loss of focus.
In Team Topologies, the team is treated as the smallest meaningful unit. And the key idea behind that is the fact that sustainable outcomes come from well-structured, collaborative teams rather than separate heroes.
Team Topologies isn’t a rigid framework or a call for company-wide reorganization. Instead, it starts with identifying value streams and understanding how work actually flows through the organization.
Thiago emphasized that meaningful cultural change within teams is neither strictly top-down nor bottom-up. It requires a combined approach. Such changes take time. It can be quarters or sometimes even years. The key is to create sufficient awareness of the broader context, the processes in place, and the value stream itself. By determining bottlenecks and reducing handovers, organizations can begin shifting toward a more collaborative and efficient culture.
This cultural shift is deeply tied to technical practices rooted in DevOps: continuous integration, continuous deployment, fast feedback loops, trunk-based development, and robust testing. But it’s not just about better tooling. It’s about bridging the gap between business and technology.
One of the biggest mindset shifts is moving away from a factory-style model where tech teams wait for perfect requirements before building. Instead, developers must become more curious about the business and more engaged with customer needs. Collaboration shouldn't be sporadic, and it shouldn’t be handed off via tickets or rigid requirements. It should be ongoing.
The core challenge lies in bridging the gap between existing infrastructure and organizational culture. It can’t be imposed top-down through mandates or principles alone.
The real test of culture comes during crises like missed deadlines, outages, or security issues. In such situations, nobody wants to take responsibility for that because responsibility can be really painful.
Often, it is not about people avoiding responsibility, but about misalignments and overloaded teams that make real ownership nearly impossible.
That’s where the principle of fast flow becomes crucial. To avoid such situations, it is necessary to reduce cognitive load, streamline knowledge requirements, and minimize distractions. This will allow teams to focus on real ownership and deliver value more effectively.
According to Thiago, there is no one-size-fits-all model for implementing cultural changes. One of the biggest challenges is building psychological safety, which is a prerequisite for any meaningful transformation. If even one person on a team doesn’t feel safe, the team as a whole isn’t truly safe.
Psychological safety starts with trust among teammates, across roles, and with leadership. Trust isn’t built through blind agreement. It’s built through transparency.
For Thiago, a practical way to foster trust is to surface assumptions and clearly explain decisions. People don’t have to agree with every call, but they should understand the rationale behind it. Disagreement is fine when it is followed by commitment and free of blame if things go wrong.
One of the most impactful ideas to emerge from Team Topologies is the concept of platform groups. They are responsible for building and maintaining internal platforms, tools, services, and building blocks that reduce the cognitive load for product-focused teams.
Thiago explained that teams that are directly delivering customer value are often overwhelmed. They are expected to handle everything: databases, deployment pipelines, infrastructure, testing frameworks, compliance, programming patterns, and business context. That’s an unrealistic cognitive burden, and it’s often why these teams default to focusing only on the technical layer.
Platform groups solve this by offering clear, reusable paths, or prebuilt ways to deploy services, manage infrastructure, or handle CI/CD. Their main goal is to streamline delivery by eliminating unnecessary friction.
However, many companies misapply this concept. They form a single overloaded “platform team” tasked with managing everything, from CI/CD and data infrastructure to Git workflows. As a result, such teams become a bottleneck themselves. That’s why the shift to true platform groups is important. Here, it is essential to keep in mind that they should be purpose-driven, focused units with clear boundaries, allowing for scale without burnout.
Thiago also highlighted another team pattern from Team Topologies. It is enabling teams. They unite cross-functional experts, such as architects or systems specialists, who embed temporarily with other teams to unblock problems, offer guidance, and enable better practices before moving on. Companies should think of them as internal consultants focused on capability-building, not control.
Cloud computing introduced a profound shift in the technology landscape.
The first big transformation it enabled was accessibility. Cloud computing removed the barrier to entry. It turned infrastructure into a utility that is available on demand based on the pay-as-you-go principle. It enabled startups and solo entrepreneurs to bring ideas to life without the need to secure bank loans just to spin up their first server.
But the second wave of transformation came with serverless technologies (or, as Thiago calls it, “serviceful” computing). Instead of managing servers or configuring infrastructure, teams can now focus almost entirely on solving business problems. These new patterns allowed developers to work faster, experiment more freely, and scale effortlessly. This approach closely aligns with the principles behind Team Topologies.
Thiago admitted that this shift was the biggest he had seen in his career before the AI transformation that we can observe today.
However, he emphasized that not everything belongs in the cloud. There are workloads that make a lot of sense to keep on-prem. This is especially relevant for companies with decades of investment, expertise, and operational maturity around legacy systems.
The real challenge of cloud transformation isn’t just technical, it’s human. Telling someone their years of expertise with data centers or custom infrastructure are no longer needed can trigger fear and resistance. That’s why change management becomes essential.
Many companies today embrace a “cloud-first” approach. But, as Thiago noted, it’s often cloud-first only until compliance or cost gets in the way.
The problems typically begin when companies attempt a “big bang” migration and try to rebuild or replatform everything at once.
Thiago recollected cases where highly competent teams are tasked with rebuilding existing systems from scratch on the cloud, but team members didn’t have enough experience in cloud-native patterns.
What comes next is often a “lift and shift” migration. It means that applications are moved to the cloud using the same designs and operational assumptions that worked on-premise. As you can understand, this method can result in multiple issues.
Sometimes a lift and shift makes sense (for example, when delaying migration would incur hardware costs or lease renewals). But that should be the exception, not the rule. Instead, Thiago advised a more incremental, wave-based approach that includes team enablement and intentional architectural planning.
The key to successful cloud-first transitions again lies in psychological safety. Companies should help people understand why the transition is happening and show how their existing knowledge can evolve in a cloud context.
Today, there are a lot of talks about cloud repatriation, which presupposes moving workloads back to on-prem. However, Thiago clarified that, in practice, he rarely sees this happening at scale.
More frequently, he can observe companies that have never completed their cloud transition in the first place. These organizations may have adopted a “cloud-first” mindset years ago, only to realize later that some workloads were better left on-prem, or that not all systems needed to move.
According to him, it’s vital to understand that not everything needs to be in the cloud.
But today, cloud providers are prepared even for scenarios that require local infrastructure. Quite often, they offer hybrid options. For instance, Outposts by AWS bring AWS-managed infrastructure into the customer’s data center, still connected to the broader AWS ecosystem. It means that businesses can maintain full control locally, but the rest of their systems can still run in the cloud.
At the same time, he also highlighted that it’s a myth that running LLMs on-prem is automatically more secure. If you are calling a third-party AI endpoint with no guarantees, that’s one thing. But platforms like AWS Bedrock give you private, VPC-based endpoints where no one else can access your data.
According to Thiago, the 80/20 rule is a good one to describe what is happening in the modern IT infrastructure. 80% of workloads can be handled by broadly available, standardized solutions, while 20% will always require specialized, often bespoke approaches.
He explained that platforms like AWS have matured to a point where the majority of business needs can be met using higher-level, off-the-shelf services. The extensive partner ecosystem has enabled businesses to build powerful platforms on top of AWS, without having to reinvent the wheel.
Most businesses no longer need to create their own data platforms from scratch. There are already high-level solutions that help them avoid most of the complexity.
However, many large enterprises still run highly customized legacy systems, often built on mainframes, and in some cases written in outdated languages with hundreds of thousands of lines of code. These systems are not easy to modernize. But they may be too critical to simply discard.
Thiago explained that the middle layer, which is the part between front-end experiences and the legacy back ends, has already been undergoing modernization for years.
What is left now is the hardest part: modernizing the base layer. It can be a real challenge, especially when companies face a knowledge drain after original developers retire.
That’s where AI and ML come into play.
AWS, for instance, provides tools like AWS Q Transform for mainframe apps. It leverages AI to analyze and explain complex legacy codebases, making them easier to understand and refactor.
The explosion of interest in generative AI and large language models has captured global attention. Nevertheless, Thiago cautioned against abandoning the foundations of traditional ML, which continue to deliver significant value across industries.
In the conversation with Max, Thiago urged organizations not to overlook the decades of progress in statistical learning, which have become overshadowed in the post-ChatGPT era. Since the launch of ChatGPT that happened in November 2022, much of the industry’s focus has shifted disproportionately toward LLMs and generative models, often at the expense of simpler and more efficient ML solutions.
Thiago compared today’s LLMs with a battalion of interns. Modern LLMs are capable of generating content, conducting research, and providing ideas, but they are inherently biased and often lacking in precision or authority.
“They speak with confidence, like white Reddit males who think they’re always right,” Thiago joked.
Hallucinations, inconsistency, and lack of source traceability are among the main issues related to the mass use of large language models. Thiago views this as a call to action for better guardrails, source attribution, and AI literacy.
Max also asked Thiago to share advice for leaders who want to implement AI solutions and build resilient technical infrastructures.
“Be empathetic and be kind. That is the most important thing that I can tell people. Everything else will follow from it,” Thiago said.
With all the changes that they can bring, technologies are just tools, people are the drivers of transformation.
Leaders must resist the urge to chase innovation for innovation’s sake. Instead, they should focus on enabling teams, simplifying processes, and creating environments where individuals feel safe, valued, and heard. This is the main conclusion that can be made from this insightful conversation.
Want to learn more about the world of business and technology? New Innovantage episodes will be available soon.
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
.png)
.png)
