Ícone do site Caixa Mágica Software

Qualigentic

Qualigentic — Hero

Agentic AI built for software quality.

The end-to-end platform that reads requirements, generates tests, executes them, and maintains your suite — on your infrastructure, on your terms.

On-premise & Private Cloud
Multi-framework
DORA · Solvency II · PSD2
Qualigentic — The Challenge

CI/CD ships faster. QA capacity does not. Regulators are now part of the conversation.

60–70%
of QA effort is spent maintaining existing tests — not writing new ones
DORA
Solvency II · PSD2 — quality is now an audit topic. Regulators want evidence chains, not screenshots.
Generic
AI ≠ QA AI
Coding assistants generate code. They do not execute, monitor, or maintain a test suite.
As ferramentas tradicionais de QA não utilizam IA de forma nativa

O scripting manual e a automação desconectada não conseguem acompanhar os ciclos de lançamento cada vez mais curtos.

Assistentes de IA genéricos geram testes sem governação

No audit trail, no project context, no integration with the QA lifecycle — and no data residency control.

Falta de rastreabilidade entre requisitos, testes e execução

Quality reporting becomes guesswork. Regulator-facing evidence is incomplete or missing entirely.

Quality teams need their own AI

Agentic, governed, and on your infrastructure — not a generic productivity tool for individual engineers.

Qualigentic — Capabilities

Six capabilities. One platform. Built specifically for the QA function.

From requirements to audit-grade evidence. Not a generic coding assistant — agentic AI specialised in software quality.

01

Requirements → Test Cases

From Jira, ALM, or spec text to executable test cases — automatically.

  • casos positivos/negativos
  • Connects to Jira · ALM · Confluence
  • Free-text input from PMs or BAs
  • Manutenção Inteligente de Suites
02

Multi-framework Generation

Production-ready scripts across all major frameworks. No proprietary runtime.

  • Selenium · Cypress · Playwright
  • Robot Framework
  • Open standards in, open standards out
03

Autonomous Maintenance

Agents detect stale, broken, and redundant tests and refactor them automatically.

  • Deteção de testes obsoletos ou redundantes
  • Self-healing scripts after code changes
  • Impact analysis on every commit
04

Coverage & Gap Analytics

Functional coverage maps, gap recommendations and risk-weighted scoring.

  • Análise de cobertura funcional
  • Gap recommendations and new scenarios
  • Deteção de redundância e duplicação
05

Intelligent Execution Analysis

Separate application failures from script failures. Pattern detection across runs.

  • Classificação de falha de aplicação vs. script
  • Identificação de padrões de falha
  • Insights para evoluir a suite
06

Audit-grade Traceability

Requirement → test → execution → result. Signed, archived, retrievable on demand.

  • Full DORA / Solvency II / PSD2 evidence chain
  • RBAC, audit logs and retention policies
  • Regulator-facing evidence in one query
Qualigentic — Comparison

Why Qualigentic?

Generic AI is a productivity tool for individual engineers. Qualigentic is a platform for the QA function.

Capacidade Generic AI Assistants Qualigentic
Generate test code from requirements Suggestion only Production-ready
Execute tests, not just write them Não
Maintain the suite autonomously Não
Multi-framework output (Selenium, Cypress, Playwright, Robot) Parcial
Requirement → test → execution → archive chain Não
Data residency / on-premise option Cloud only On-prem available
DORA / Solvency II / PSD2 audit evidence Não
Pricing model Per-token (developer seat) Platform + services

ChatGPT, Claude direct, GitHub Copilot, Gemini Code Assist suggest code. They do not own the QA function.

Qualigentic — Architecture

Open standards in. Open standards out. Customer infrastructure throughout.

Multi-model AI support, agentic orchestration and deployment options to fit every context — including regulated industries.

Entradas
Requisitos funcionais
Histórias de utilizador / Especificações
Testes existentes
Jira · ALM · Confluence
Free-text from PMs
Qualigentic Platform
Motor RAG
Project context, test patterns, organisational memory
Agentic Orchestration
Multi-step planning, tool use, human-in-the-loop checkpoints
Multi-LLM Layer
Llama · Mistral · Anthropic · Azure OpenAI · AWS Bedrock
Fine-Tuning (PEFT / LoRA)
Customer-specific adaptation. Weights stay on customer infrastructure.
Governance & Audit
RBAC · Signed evidence chain · Retention · DORA-aligned controls
Integração CI/CD: GitHubGitLabAzure DevOpsJenkinsBitbucket
Saídas
Scripts Selenium
Scripts Cypress
Scripts do Playwright
Robot Framework
Coverage reports
Audit evidence

No local

Your data centre. Open-source self-hosted models (Llama, Mistral). No data egress under any condition. Recommended for banking, insurance, healthcare and government.

Private Cloud

Your tenant. Azure AI Foundry, AWS or GCP — customer-owned. Bring-your-own model and keys. Region pinning for EU, US, JP data residency.

SaaS

Caixa Mágica managed. Fastest time-to-value for non-regulated workloads. EU-hosted, SOC 2-style controls, Anthropic / OpenAI / Azure OpenAI selectable.

Tiering is by capability, not by deployment. All tiers support on-premise, private cloud and SaaS. Regulated clients can start on-premise from day one — no lock-in to a cloud-only path.
Qualigentic — Packages

Tiers by capability, not by deployment.

From first AI adoption in QA to mission-critical operations in regulated industries. All tiers support on-premise, private cloud and SaaS.

Entrada

QA teams adopting AI for the first time

  • Geração de testes para múltiplas estruturas
  • Requirements → executable scripts
  • Dashboards e relatórios
  • Single-team scope
  • Any deployment model
Saiba mais
Empresa

Mission-critical operations, regulated industries

  • Tudo em Profissional
  • Full DORA / Solvency II / PSD2 audit chain
  • Fine-tuning (PEFT / LoRA) on customer infrastructure
  • SLA + dedicated success engineering
  • Co-created roadmap
  • QA / TestOps consulting
Saiba mais

Regulated clients can buy Starter on-premise from day one. Commercial flexibility is part of the offering.

Qualigentic — FAQ

Perguntas frequentes.

How is Qualigentic different from a generic AI coding assistant?

Generic AI assistants like ChatGPT, GitHub Copilot or Gemini Code Assist suggest code — they do not execute, monitor or maintain a test suite. Qualigentic is an agentic platform built around the QA function: it reads requirements, retains project context through RAG, generates production-ready tests, runs them, analyses failures, and maintains the suite — all with a governed audit chain.

Which frameworks are supported?

Selenium, Cypress, Playwright and Robot Framework are first-class. There is no proprietary runtime — all generated scripts use open standards. Additional frameworks can be enabled via the platform's adapter layer in the Enterprise tier.

Can Qualigentic run fully on-premise?

Yes — and this is available from the Starter tier, not just Enterprise. On-premise deployments use open-source self-hosted models (Llama, Mistral and others). Fine-tuning weights (PEFT / LoRA) never leave the customer's perimeter. No data egress under any condition.

How does Qualigentic handle regulatory compliance requirements?

The platform produces a signed audit evidence chain from requirement through to archived result, designed against DORA Articles 6 & 9, Solvency II Pillar 2, and PSD2 Article 95. Enterprise-tier customers get full RBAC, retention policies, and regulator-facing evidence retrievable on demand.

What does a typical pilot look like?

A time-boxed 6–8 week engagement: one application, one framework, one defined success criterion. Weeks 1–2 cover discovery and environment setup; weeks 3–5 run the first generation and execution cycles; weeks 6–7 validate coverage and produce a regulator-facing evidence sample; week 8 delivers an executive readout with ROI calculation and expansion plan. The pilot fee is credited 100% against the year-one platform fee on conversion.

Which AI models does Qualigentic use?

Qualigentic is model-agnostic. Supported options include open-source self-hosted models (Llama, Mistral), Anthropic Claude (preferred partner), Azure OpenAI / OpenAI, and AWS Bedrock. Customers choose the provider that fits their compliance scope — no Qualigentic lock-in.

Does Qualigentic replace QA engineers?

No. It removes the manual scaffolding, maintenance and triage work — freeing engineers to focus on test strategy, complex scenarios and quality engineering. Humans approve, escalate and override at every step of the agentic loop.

Qualigentic — Get Started

Agentic AI for Software Quality.
Vamos conversar.

Personalised demo, no commitment. We'll show how Qualigentic fits your context — frameworks, infrastructure and governance. Pilot fee credited 100% on conversion.

Sair da versão mobile