Generative AI is accelerating software development across banking and financial services, but its ability to deliver real innovation depends on whether quality engineering can keep pace.
As banks push more AI-generated code and models into production, testing, governance and trust are becoming decisive factors in how fast GenAI can be deployed safely.
In highly regulated environments, quality engineering is moving from a back-office function to a core enabler of AI adoption. Without stronger automation, continuous testing and tighter controls, GenAI risks increasing operational and compliance exposure rather than reducing it.
Dror Avrilingi, Head of Quality Engineering, Data and GenAI Studios at Amdocs, said GenAI’s promise in financial services will only be realised if testing evolves alongside development.
“GenAI is accelerating innovation in financial services, but only if Quality Engineering keeps up,” Avrilingi stated, describing quality engineering as “an unexpected innovation accelerator when paired with next-gen, agentic Quality Engineering.”
While GenAI is “driving change at astonishing speed” across industries, Avrilingi observed that “progress is slower in highly regulated industries, such as financial services.”
“GenAI is accelerating innovation in financial services, but only if Quality Engineering keeps up.”
– Dror Avrilingi
Many banks are already using GenAI in software development, and Avrilingi pointed to research showing that “GenAI boosts developer productivity by 20–50%.”
Faster development, however, does not automatically translate into safe innovation. “In a world where innovation is easier, quality engineering (QE) becomes more than a trust, technical, and compliance checkpoint,” Avrilingi said.
“QE serves as a critical accelerator or enabler, especially in highly regulated industries,” he stressed.
According to Avrilingi, banks must balance speed with regulatory and operational discipline.
“Financial services institutions must balance the desire for rapid customer-facing innovation with the need to maintain compliance with financial regulations, and guarantee accuracy and security,” he said.
“QE is essential to meeting compliance requirements and maintaining trust within customer experiences that include GenAI.”
He also warned that GenAI introduces new quality risks if testing does not scale. “As much as 40% of GenAI-generated code requires remediation,” Avrilingi said.
“When you can produce code almost instantly, you need similar automation in quality processes to streamline remediation and catch errors likely to cause issues in production.”
Synthetic test data
Embedding GenAI directly into quality workflows is central to that shift. “GenAI goes beyond ML by assisting you in automating test generation for specific applications and a diverse range of scenarios,” Avrilingi argued.
He added that “GenAI changes the game by rapidly creating regulatory-compliant synthetic test data suited to specific use cases.”
Avrilingi urged banks not to delay modernising their testing strategies. “When should you begin your GenAI QE journey? Now,” he said.
“Your use of GenAI in QE needs to grow as fast, or faster, than your appetite for AI-powered innovation.”
As large language models are embedded into financial products, quality engineering is becoming a key AI governance mechanism.
“Without the proper guardrails, LLMs don’t meet FSI requirements for accuracy and reliability,” Avrilingi wrote.
“QE can help optimise LLMs by applying a comprehensive and continuous approach to fine-tuning capabilities.”
He added that “customer-ready GenAI requires testing focused on context understanding, content evaluation, and continuous feedback,” and that “a non-stop feedback loop helps the model adapt, improve, and avoid errors and bias.”
Although GenAI-enabled quality engineering remains early-stage for many firms, Avrilingi said early adopters are already seeing tangible gains.
“GenAI-augmented QE is still emerging within most FSIs, yet the businesses moving early are surging ahead,” he wrote, citing results including “a 33% decrease in testing design time,” “a 50% improvement in test coverage optimization,” and “a 60% decrease in testing certification time.”
Looking ahead, Avrilingi said autonomy will define the next phase. “The next step is Agentic AI, cognitive quality agents capable of executing tasks, adapting to change, and optimizing pipelines with increasing autonomy,” he shared.
In that model, “QE shifts from automation to intelligent orchestration: humans define the guardrails, while agent-driven systems handle scale, regression, and continuous validation at speed.”
For regulated financial institutions, Avrilingi said quality will ultimately define AI leadership. “In financial services, leadership in AI will be defined not by novelty, but by resilient performance, accuracy, and trust,” he said.
“Agentic AI represents the path to that standard, enabling quality to stay resilient and adaptive as innovation cycles compress, and to evolve as quickly as the intelligence driving development itself,” Avrilingi concluded.
COMING IN 2026
Why not become a QA Financial subscriber?
It’s entirely FREE
* Receive our weekly newsletter every Wednesday * Get priority invitations to our Forum events *
REGULATION & COMPLIANCE
Looking for more news on regulations and compliance requirements driving developments in software quality engineering at financial firms? Visit our dedicated Regulation & Compliance page here.
READ MORE
- Trust, not speed: Why AI governance is now a testing battleground for banks
- NatWest’s AI trade finance overhaul opens new chapter for QA teams
- Banking UAT moves beyond sign-off as QA takes centre stage in system rollouts
- Citi ramps up AI-driven testing in race to modernise legacy systems
- Lloyds, HSBC and NatWest get OpenAI access amid mounting concerns
QA FINANCIAL PODCASTS
LISTEN TO OUR EXCLUSIVE CONVERSATIONS
WATCH NOW
