Abstracta has introduced Quality Intelligence, an approach to software quality engineering that combines AI agents, technical and business context, operational data, and human oversight to help organizations assess software changes before release. The company is targeting banks and other enterprises that need to accelerate software delivery without losing visibility into defects, system dependencies, and operational risk. Its approach is delivered through Abstracta Intelligence, an enterprise platform built on the open-source Tero agent framework.

Abstracta has introduced Quality Intelligence, a software engineering approach designed to help enterprises understand the risks associated with software changes as AI accelerates development. The company is positioning the practice for organizations running business-critical systems, including banks, where release decisions can affect financial transactions, sensitive data, and customer services.

The announcement was made by Abstracta co-CEO Sofía Palamarchuk during the opening keynote at Quality Sense Conf 2026 in Montevideo, Uruguay. The company also presented the first version of its Quality Intelligence Manifesto, setting out a framework that connects software systems, operational context, and human judgment.

From Software Testing to Release Intelligence

AI-assisted development can increase the volume and speed of code production, but faster development does not automatically make software safer to deploy. Engineering teams still need to understand how a change affects existing functions, whether tests cover the relevant areas, and whether production incidents or historical defects indicate additional risks.

Abstracta calls the disconnect between AI-generated output and the information required for safe release decisions the “Quality Intelligence Gap.” Rather than treating quality assurance as a final testing stage, its approach brings together requirements, tickets, code, tests, incidents, and production behavior to help teams evaluate changes throughout the software delivery process.

The distinction matters in banking and financial services. Core banking platforms often combine modern applications with older systems and complex business rules. A software change that appears small in isolation can affect connected workflows, making it difficult for teams to assess release risk from test results alone.

How Abstracta Intelligence Works

Abstracta puts the approach into practice through Abstracta Intelligence, an enterprise AI platform built on Tero, its open-source framework for developing context-aware agents for quality assurance and software delivery. The platform is designed to connect with existing enterprise tools rather than requiring organizations to replace their entire development environment. Its integrations include Jira, continuous integration and delivery pipelines, and legacy platforms such as AS/400 and core banking systems.

The company describes several functions within the workflow. System Understanding draws together documentation, tickets, code, and tests to explain how a feature or business rule operates. Change Impact identifies modifications and potentially affected areas, while Risk Assessment considers historical defects, test coverage, and incidents.

Release Readiness then consolidates available evidence to support a go/no-go decision. Knowledge Capture aims to turn specialist knowledge into reusable context that teams can apply across future work. Abstracta also offers dashboards covering testing maturity, DORA metrics, and release readiness, giving engineering, quality assurance, and leadership teams a shared view of delivery performance.

Human approval remains central to the model. Abstracta says its approach includes defined roles, audit trails, and human sign-off for release decisions. That distinction is important: AI agents can help gather evidence and identify possible risks, but the platform’s stated workflow does not make the software itself the final authority on whether a release should proceed.

Financial Services Risk and Governance

The launch comes as financial institutions develop more formal approaches to AI risk management. In February 2026, the U.S. Department of the Treasury released its Financial Services AI Risk Management Framework, which adapts broader AI risk-management principles to financial-sector operational, regulatory, and consumer-protection concerns. The framework emphasizes evaluating AI use cases and managing risks across the AI lifecycle. <Cite refs={[“turn953687search0”]}/>

Abstracta says its Quality Intelligence model shares principles such as accountability and risk-aware adoption. The company also describes a five-stage maturity model intended to help organizations assess their progress from individual experimentation toward AI-assisted release decisions. The maturity model is Abstracta’s own framework; it should not be confused with a regulatory requirement or an official Treasury standard.

Reported Results Need Context

Abstracta says its work across customer engagements has produced a 40% reduction in production defects in under six months, regression cycles reduced from days to about 30 minutes, and more than 2,000 hours saved per month through AI agents built on Tero. These are company-reported results. The announcement does not provide enough project-level detail to independently assess the baselines, measurement methods, or extent to which those outcomes can be reproduced across other organizations.

The company points to nearly two decades of software quality work and projects involving global banks such as BBVA and Santander. It also describes work with Bantotal’s research and development team and Microsoft’s AI Co-Innovation Lab on a generative AI copilot designed to help teams understand a large core banking system.

For banks considering AI in software engineering, the broader question is not simply whether agents can speed up testing. It is whether teams can integrate AI into established controls, trace its outputs to useful evidence, and retain accountability for production decisions. Abstracta’s Quality Intelligence proposition addresses that need, though its real-world value will depend on implementation quality, integration with existing systems, and independently measurable outcomes.

Market Landscape

Financial institutions face a dual challenge: modernizing legacy software and adopting AI-assisted development without weakening operational controls. Quality engineering platforms are increasingly being positioned as a way to connect fragmented development information, automate repetitive QA tasks, and give teams better visibility into release readiness.

Abstracta’s differentiation is its focus on contextual risk assessment and human-approved release decisions, rather than test automation alone. Its proposition will need to be evaluated against practical factors such as integration complexity, evidence traceability, security requirements, and whether reported productivity improvements translate into fewer production incidents.

The U.S. Treasury’s Financial Services AI Risk Management Framework provides relevant industry context for institutions establishing AI governance processes, although it does not endorse Abstracta’s product.

Top Insights

• AI-assisted quality engineering: Abstracta connects software context, historical defects, tests, and production signals to help teams evaluate changes before release.
• Banking-focused workflows: The platform is designed to work with existing tools and legacy systems, including AS/400 and core banking environments.
• Human accountability: AI agents prepare evidence and risk assessments, while people retain responsibility for final release approvals.
• Company-reported outcomes: Abstracta claims fewer production defects and shorter regression cycles, but the announcement does not independently verify these results.
• Governance remains essential: Financial institutions must assess how AI-generated recommendations are documented, reviewed, secured, and integrated into existing risk controls.

Get in touch with our fintech expert