Why We Launched RE:WORK

Before RE:WORK began, we conducted an internal survey. It showed that most employees were already experimenting with AI tools, but only a small percentage reported meaningful time savings. The conclusion was clear: informal usage does not create consistent or measurable improvements. A structured approach was needed.

As the transformation progressed, two internal publications helped describe the broader context of what we were observing.

Our CSO wrote about how AI affects cost, time, and value delivery:

→ Why at Polcode, We Treat AI as a Strategic Layer, Not Just a Tool

Our CTO explained why engineering requires structured processes to make AI useful rather than unpredictable:

→ AI Won’t Fix Your Engineering Process. It Will Expose It

Both texts reflect what we are seeing in practice: AI becomes effective only when the underlying processes are stable and uniform and consciously designed.

Where We Are Now: Onboard Phase in Progress

The Onboard phase focuses on readiness and adoption, following a structured implementation approach:

1. RE:WORK Team Formation. A dedicated RE:WORK Team, representing all key specialisations across the company, was established. This team meets regularly to coordinate the transformation, prioritise AI&Automation Workflows, and ensure smooth organisational change. Internal AI Advocates were appointed to support teams daily, answer questions, gather feedback, and help develop early best practices. Their role is critical in keeping adoption practical rather than theoretical.

2. Identification of Repetitive Tasks. Each role identified 3-5 repetitive, low creativity, high-frequency tasks suitable for automation. The selection was based on clear criteria:

   • high frequency, time intensity, repetitiveness, low creativity,

   • documentation load and measurable cost impact.

   These tasks form the foundation for our AI & Automotion Workflows.

3. Tool Selection. Based on the identified tasks and role requirements, appropriate AI tools were selected for each specialisation. The selection process considered functionality, integration capabilities, and alignment with identified workflow needs.

4. Legal and Safety Review. From Day 1 of RE:WORK, safety was treated as a non-negotiable foundation, not a later checkpoint.

   Every AI tool undergoes a rigorous legal and safety review before teams are granted access. These reviews are not treated as one-time approvals. AI tools, their capabilities, and even their terms of service evolve constantly, which is why we continuously monitor changes in tooling, data-processing rules, and legal frameworks.

   Non-negotiables:

   1. clear ownership of inputs and outputs,

   2. no model training on business data by default,

   3. enterprise-grade security standards,

   4. predictable data-processing rules,

   5. GDPR-ready configurations,

   6. organisational control (business-grade licenses only; no private or individual accounts)

   Additionally, AI tools are used only with explicit client approval.

   These standards form part of a broader AI Policy, which will be made publicly available.

5. Tool Access Distribution. Every role received access to company-funded AI tools along with clear guidelines for safe and compliant use. Teams were also supported with practical onboarding materials: how to start, where AI fits into daily work, and where human responsibility remains mandatory.

6. AI&Automation Workflows Development, Testing, and Human-in-the-Loop. Based on the analysis of repetitive tasks, we created the first set of AI&Automation Workflows designed to streamline work across roles. Some of them are already being tested by teams. Additional workflows are in active development and are planned for release to teams in the upcoming weeks.

The Onboard phase continues. Adoption is growing, and feedback loops are now in place. This stage is expected to conclude around early 2026, transitioning into the Operationalize phase once the core workflows are validated.

What We Are Learning During Onboard

A key insight of Onboard is that AI is not an add-on. RE:WORK is fundamentally reshaping how we work and deliver projects. We are redesigning workflows, documentation patterns, and decision-making structures to make AI a stable, predictable part of our delivery process.

The second insight is that standardisation is essential for automation to work. AI does not fix inconsistencies; it reveals them. As teams began using AI daily, they quickly identified gaps in documentation, processes, and expectations. Standardisation is now treated as a prerequisite for reliable AI adoption - a perspective directly aligned with our CTO article.

The third area is responsible use. All AI tools used at Polcode go through legal and security reviews. Business-grade licenses are mandatory; private accounts cannot be used for project work. Teams follow structured rules for data handling, including when working with regulated industries. These standards form part of a broader AI Policy, which will be made publicly available.

What Happens Next

Once Onboard concludes, the program moves into Operationalize - testing AI workflows across the entire company, gathering insights, and refining them based on real-world usage. This phase focuses on building AI maturity organisation-wide. The final Embed phase will standardise these practices across all roles and projects, making AI-supported delivery repeatable and predictable at scale.

As the transformation progresses, we will continue sharing updates. Each stage builds on the previous one. The objective remains consistent: make delivery more predictable, reduce manual work, and strengthen the foundations of software development across the company.

Closing Remarks

RE:WORK is an ongoing program. At its core, RE:WORK is about freeing specialists from repetitive, low-creativity tasks so they can focus on what truly matters - solving clients’ most complex challenges and delivering innovative solutions. By giving teams back time for creative, analytical, and strategic work, we are strengthening the foundation of how software is built at Polcode.

The transformation continues, and we will report further developments as the next phase begins.