AI Insights

The development of AI tools is adjusting traditional development processes, requiring new skills and flexibility from specialists. At Godel, this direction is supported through internal educational programmes. We spoke with course leaders Aliaksandr Haurylau and Tatsiana Toustsik about how training is structured today and how these tools are impacting team roles.

Internal AI courses are currently being held at Godel. What was the catalyst for creating this training?

Alexandr: First and foremost, it’s market challenges. There is a global transformation happening across the entire IT industry. Thanks to AI, work productivity has increased many times over, and it’s important for us to offer our clients a new level of service. Our courses respond to current trends and offer every employee the opportunity to master tools that qualitatively change familiar processes.

Which programmes are currently available, and how has their direction changed?

Alexandr: Currently, I’m directly overseeing the development of the AI First SDLC course. We also have the AI Enable course. We started developing that programme last autumn; it’s dedicated to working with Copilot and similar tools. The last groups are completing their training now. This course is available in our programme portfolio and is offered both internally and to our clients.

Tatsiana: A lot has changed since the start of development. Initially, we planned to create a course on SDD (Spec-Driven Development), but during our experiments, we found that the tools are still quite raw. Technically, it’s currently impossible to have a single universal solution that covers the tasks of all roles on a project. Therefore, we changed direction and renamed the course – it’s now called AI First SDLC.

What does the AI First SDLC concept entail?

Tatsiana: The key here is the acronym SDLC (Software Development Life Cycle). The core idea is that the boundaries between roles are blurring significantly. At different stages of the development cycle, tasks get ‘blurred’: a developer might partially take on the functions of a business analyst, while an analyst, with the help of AI, can implement a task they’ve described well. This course will be unified for the entire company. We want everyone to understand not just their own role, but also how to work at the intersection with related specialities. The boundaries between roles will be blurred – in fact, they are already blurred.

Alexandr: The AI First course is in the final stages of development. It will differ from classic programmes: it’s a self-learning format. Our goal is to introduce everyone to the possibilities of AI, specifically the agentic approach to work. We provide the foundation so that specialists can then independently study specific issues using modern tools and their own experiments.

Who is involved in the development, and who will be able to take the training?

Tatsiana: The team is truly cross-functional. Absolutely all departments are involved: Software Engineers, ADC, BA, DevOps, Designers. It’s a massive collaborative effort across the whole company.

Alexandr: The idea is that everyone, without exception, should go through a unified programme. Everyone must be ready to perform any role within the SDLC. Of course, a business analyst isn’t required to become a deep DevOps engineer. However, practice shows that, by leveraging AI agents’ capabilities, a business analyst can quite easily handle a task at the POC or MVP level on their own and clearly convey their vision for the product to the team or client. This makes the development process as transparent and fast as possible.

Why should specialists take on additional roles?

Alexandr: We’ve moved away from the concept of T-shaped specialists and transitioned to an M-shaped model, where employees can flexibly switch between tasks depending on the situation. Coding has ceased to be the exclusive prerogative of developers, and analytics is no longer the exclusive domain of business analysts. There are now enough tools that allow, for example, a developer to independently handle a client’s requirements without intermediaries. In modern teams, role boundaries are becoming more flexible. In the future, clients will increasingly look for comprehensive services in which it’s less important who exactly performs a role at any given moment.

Tatsiana: AI allows you to perform someone else’s job at a very decent level without diving into the enormous depth of the specifics. This doesn’t mean that roles will disappear, but AI can handle routine tasks independently when colleagues are busy with more important matters. For example, a business analyst understands the context perfectly, and if you transfer that context to an AI tool, the result is working code.

What are the main advantages of this approach for clients?

Alexandr: Speaking of the benefits for clients, three key factors stand out. First is resource efficiency: our target benchmark is to achieve a balance in which up to 80% of tasks are solved using AI resources, while 20% rely on expert human participation. This allows us to significantly optimise the process and increase the value of our offer.

Second, Time to Market has changed completely – development speed with AI agents is now simply incomparable to traditional methods. And finally, scalability: with AI, the onboarding process is significantly simplified. For example, technology allows code to be translated into clear documentation, so new specialists can instantly get answers about how even the most complex systems work.

Tatsiana: The whole world is talking about how the performance of such teams is growing. From the perspective of quality and approaches, the process moves faster than if we were writing everything purely ‘by hand’.