Embedding AI Across the Development Process: Inside Cywareness' Warsaw Workshop

The Cywareness R&D team recently gathered in Warsaw for a three-day workshop focused on embedding AI across the development process. Held as part of CyberproAI's broader AI transformation, the workshop examined how AI can become an integral part of the way teams plan, build, validate and continuously improve their work.

Over the three days, the team turned two weeks of preparation into practical AI capabilities built around its real workflows, codebase and day-to-day challenges.

Embedding AI Into the Development Process

The workshop had a clear purpose: to move from treating AI as an external tool to embedding it in how the team works.

Rather than introducing a standard toolset, the team developed personalized AI agents that understand each member's responsibilities, methods and requirements.

The team also established shared practices for AI-assisted planning, implementation and validation. Completed work had to be run and tested, recurring solutions were captured as reusable skills, and the shared "Superpowers" library would keep improving over time.

"We are moving from building products to building a machine that builds products." — Ofry Ram, Head of Tech at CyberproAI and CTO of Cywareness

Building Around Real Workflows

Two weeks of preparation preceded the workshop, giving the team time to understand how developers actually work in practice.

The team analyzed developers' interactions with Claude, reviewed GitHub commit history and identified recurring tasks, challenges and working patterns across the development lifecycle.

The aim was to build around the team's actual workflows, codebase and day-to-day needs rather than assumptions.

Creating a Shared Foundation

Those findings shaped everything that followed.

During the workshop, participants refined their personal agents and created skills for specific development activities. Each agent was designed around its user while contributing to a shared foundation that benefits the entire team.

Together, these skills form a library that can expand as the team gains experience. Capabilities developed through one person's work can be captured, refined and reused across R&D, allowing individual knowledge to strengthen the wider organization.

The next phase will focus on adapting the development process and migrating knowledge and data to GitHub, creating a synchronized foundation across the implementation organization.

Insights from each sprint will inform workflows, agents and the skill library, supporting faster delivery, fewer blockers and quality built into the process.

From Individual Tools to Shared Capability

AI use is often fragmented: different prompts, agents and workarounds. Cywareness saw an opportunity to bring those individual approaches into the open.

A prompt one developer has refined may solve a colleague's problem; an agent built for one recurring task may save the entire team hours. Capturing those solutions in a shared library turns isolated experimentation into a team capability.

Just as importantly, the workshop gave the team a chance to work face to face, spend time together and strengthen the relationships behind the product. Shared tools are most effective when they are supported by shared habits, trust and a willingness to keep learning together.

The workshop reflects CyberproAI's broader approach to AI transformation: combining technology with practical changes in how teams work, learn and deliver.

The Warsaw workshop marked an important step in embedding AI across Cywareness' development process and establishing a shared foundation the team can continue to strengthen through daily work.