Polytech Software
Talk to us

Independent QA Partnership and AI-Driven Quality Transformation

Long-term quality ownership since 2009 — manual + automation, end-to-end control, and AI-enabled operations

Read More

Context

Our cooperation started in 2009. Since then, we have acted as an independent quality authority for the product — not just "checking boxes," but owning critical end-to-end paths, catching regressions before release, and keeping a consistent quality bar across multiple teams.

The client is a company building a cloud platform for storing, editing, and distributing print marketing materials. International brands provide access to templates for local offices and distributors, after which materials are produced by local print houses or delivered to recipients. The platform supports assets such as brochures, posters, and event materials — which makes stability, correctness, and predictable releases business-critical. Among notable end users are BMW, Lufthansa, Luxoft, Hexagon AB, and Miele.

The product is developed by a distributed engineering organization spanning 12 countries. In practice, we often serve as the "glue" between business, engineering, and operations: turning requirements into testable artifacts, closing gaps in user scenarios, and advising where automation delivers the highest ROI.

Key Conditions

  • 1

    Scope of quality ownership

    • Manual testing and automated tests
    • Third-party integrations and APIs
    • File-processing pipelines for print production
    • Quality of final printed output (where applicable)
    • Payment scenarios inside ordering flows
  • 2

    Quality mandate

    Protect critical end-to-end journeys, prevent regressions, and maintain a unified quality bar across teams and release streams
  • 3

    Operating model

    Independent QA function working alongside multi-team delivery and operations

Challenge (Why It Was Hard)

The challenge wasn't only technical complexity — it was system complexity:

  • Multiple teams across countries with different delivery cadences
  • End-to-end flows spanning integrations, file pipelines, production constraints, and payments
  • High cost of failure (regressions affect real customer operations, printing, and ordering)
  • The need to keep quality consistent while the organization and tooling evolved over time

Turning Point: 2021 and the AI Era

In 2021, AI became mainstream, and the client decided to rethink how teams operate and how work gets done faster with fewer bottlenecks. Formally, the QA department was reduced — but instead of a step backward, this became a growth point.

We proposed and executed a pragmatic transformation:

  • A group of strong QA engineers with automation background transitioned into engineering roles within the new Delivery Platform, keeping quality "inside the code" and delivery mechanics
  • Another group, embedded AI into product and operational workflows to reduce routine work and produce faster signals about risks and regressions

As a result, the Platform quickly evolved into an innovation engine — while quality coverage and accountability stayed intact.

Our Approach

We treated quality and delivery mechanics as an operating system:

  • Keep independent QA as the quality benchmark for critical flows
  • Move part of quality ownership into engineering (quality-in-code) where it scales best
  • Use AI to reduce operational friction and accelerate risk detection (faster signal, less noise)
  • Standardize playbooks so teams can execute consistently, regardless of geography

What We Delivered

  • 1

    Independent end-to-end QA ownership

    Continuous validation of critical user journeys, regressions caught before production releases, consistent acceptance criteria across teams
  • 2

    Automation strategy and execution

    Focusing on high-value flows (integrations, APIs, payment paths, and print pipeline checkpoints)
  • 3

    AI-enabled operations toolset

    • Slack-based bots integrated with knowledge and service metadata (Confluence + Service Catalog)
    • automated release approval workflows
    • AI hints and operational alerts integrated into incident/ops processes (e.g., OpsGenie)
  • 4

    Modern stack enablement

    We led selection and onboarding of a contemporary dev + test toolchain aligned with the new Delivery Platform goals
  • 5

    Playbooks for engineers and product managers

    From technical workflows to the practical rules of writing crisp, testable tasks and requirements
  • 6

    Continuous tech scouting

    Evaluating new tools, selecting what actually helps, and introducing it without disruption
  • 7

    Service Catalog as the backbone

    • automatic synchronization with collaboration and delivery systems (Slack/Jira/HR systems)
    • a single, always-current view of services, owners, and dependencies
    • reduced operational friction, faster onboarding, and more predictable releases

Technology Stack / Tools

Engineering / runtime

REST APIs, Rebus, MS SQL, Azure Service Bus, Kubernetes

Collaboration & delivery

Jira, Confluence, Miro, Slack, OpsGenie, Service Catalog

Practices

Agile/Scrum, test automation, QA gates, AI integrations into product and ops workflows

Results / Client Outcomes

  • Independent QA remained the quality bar for critical end-to-end paths while the organization shifted toward quality-in-code
  • Reduced operational friction through standardized playbooks and a shared service/dependency view
  • Faster feedback cycles and earlier risk visibility via AI-enabled alerts and automated workflows
  • More predictable releases and smoother onboarding across a distributed multi-country engineering organization
  • The QA team budget was reduced.