// how we work

Collaboration and delivery

How independent companies collaborate and deliver shared projects.

// between companies

Working agreements

Participants agree the terms of each initiative. Each company determines how to organise its own business. Roles, success criteria and quality requirements are agreed before work begins.

Roles and coordination

The lead company, product and technical owners, contacts and the decision-making process.

Commercial terms

Scope, payments, client relationships, change management and the process for resolving disagreements.

Rights and resources

Ownership of outputs, confidentiality, licences, data access and rules for component reuse.

// delivery model

How our work is organised

Competency-Driven Management

Project teams are engaged based on domain expertise, not hierarchy or resource availability. This eliminates the handoff effect between contractors.

Single Accountability Point

Each project has an agreed lead member company and client contact. The contractual model, SLA and reporting are defined around the participating teams and their responsibilities.

Shared Technology Foundation

UnityBase and other ecosystem platforms have their own architectures. Compatibility is established through agreed APIs and data contracts, identity management and integration checks.

Security by Design

Threat modelling, least privilege, data protection and recovery requirements inform architecture. Regulatory obligations depend on the data, sector, jurisdiction and each project participant’s role.

// from task to operations

Delivery stages

  1. Feasibility

    Define the task, baseline, expected value and success criteria. Compare AI with rules-based automation.

  2. Data and architecture

    Identify sources, data quality, permissions, integrations, deployment environment and component owners.

  3. Pilot and evaluation

    Test quality on representative cases, errors, security, latency and the full cost of a correctly completed operation. Use the result to decide whether to scale.

  4. Production launch

    Test load and recovery, configure monitoring, train users and agree support.

  5. Operations

    Monitor quality, costs and incidents. Changes to models, prompts, data sources or tools require evaluation and a controlled release.

// requirements

Standards & compliance

Standards, regulatory requirements and engineering practices serve different purposes. Applicability is documented per project; a company’s certification is established by a valid certificate and its scope.

ISO/IEC 27001:2022
Information Security

ISO/IEC 27001:2022 covers information security management, risk assessment and controls for access, incidents and suppliers. Implementing controls and obtaining independent certification are separate steps.

NIS2
EU Cybersecurity

NIS2 applies through EU Member States’ legislation to defined entities and sectors. Financial institutions additionally require assessment of DORA and ICT supplier obligations.

GDPR / Data Act
Data Protection

GDPR governs personal data. The Data Act separately addresses access to connected-product data and switching cloud providers. Access rights, portability and contractual terms inform architecture.

Zero Trust
Security Architecture

Access is evaluated using identity, device posture and request context. Least privilege, segmentation, MFA and workload identity controls complement network protection.

DevSecOps
Development & Security

A secure lifecycle informed by NIST SSDF and OWASP ASVS 5.0 verification: code, dependency and secret analysis, SBOMs, artifact provenance and vulnerability remediation.

ISO 9001:2026
Quality Management

ISO 9001:2026 is the current quality management edition. Transition from the previous edition is planned with the certification body; each company’s status is checked separately.

ISO/IEC 42001 / AI Act
AI Governance

AI risk management, accountability, data quality and human oversight. AI Act applicability depends on intended use and organisational role; ISO/IEC 42001 certification does not replace regulatory obligations.

Cyber Resilience Act
Digital Product Security

For products within CRA scope: vulnerability handling, updates and assurance evidence. Reporting obligations apply from September 2026 and the main requirements from December 2027.

Ukraine
Information Protection

Security authorisation and integrated information protection systems (KSZI) under current and transitional procedures. Protection, cryptography and assurance requirements follow the information category and system.

// engineering practices

Approaches to development

AI-assisted development

Approved tools and data rules, small changes, tests and accountable review. Measure impact through delivery time, defects and release stability.

Agent security

Evaluation informed by NIST AI RMF and OWASP guidance for LLMs and agents. Tools have limited permissions, spending limits, action logs and defined human approval points.

Agile / Scrum / Kanban

Scrum for iterative product work and Kanban for continuous flow. Cadence, work-in-progress limits, demonstrations and retrospectives are selected for the team’s context.

CI/CD

Automated builds, tests and security checks; small releases, controlled rollout and rollback. AI-assisted changes undergo the same checks and accountable review.

Domain-Driven Design

Model business domains, record decisions in ADRs and verify integration contracts. Select a modular monolith or microservices based on domain boundaries, scaling needs and operational readiness.

Infrastructure as Code

Terraform, Ansible and other IaC tools provide versioned environments. GitOps, policies and secrets have explicit controls; recovery is verified through exercises.

SRE and Observability

OpenTelemetry for metrics, logs and traces; SLIs/SLOs, error budgets, incident reviews and practical RPO/RTO verification.

Controlled AI Lifecycle

Versioned models, prompts and evaluation sets; permission-aware RAG, prompt-injection checks, restricted tools, budgets and agent autonomy. Failures have a defined safe fallback.

Platform Engineering and FinOps

Team self-service through verified templates, supported versions and transparent costs. Platform value is assessed through developer experience and service outcomes.

// after launch

Support and accountability

Support is provided by the designated member company. Service targets and responsibilities are contractual and specific to the solution.

Availability and response

The SLA defines availability measurement, support hours and the responsible company’s response commitments.

Reliability and recovery

SLOs, RTO and RPO are agreed according to system criticality, architecture and recovery test results.

AI quality and cost

AI functions have agreed quality thresholds, autonomy limits, budgets and a process for handing work to a person.

Plan together

Tell us about your company, product and expertise. Let’s explore how we can strengthen each other and identify a first shared task.