From 2014 onward, DBS Bank pursued one of the most publicly documented engineering transformations in incumbent banking. The figures the bank itself published — a 10× increase in monthly deployment cadence, 100% automated releases on the modernised stack, and AI deployment cycles reduced from 18 months to less than 5 — are the clearest answer in the public domain to a question many financial-services boards are now asking: what does it actually take for a regulated incumbent to operate at digital-native engineering velocity?
The benchmark: 10× release cadence on regulated infrastructure
Beginning around 2014, Singapore's DBS Bank pursued one of the most publicly documented engineering transformations in incumbent banking. Branded internally as "GANDALF" — an aspiration to operate at the engineering standard of the world's leading technology companies — the programme set explicit, measurable transformation targets through 20181.
According to materials presented by the bank itself, those targets included: a 10× increase in monthly deployment cadence, 100% automated application releases on the modernised stack, and a 7.5× increase in automated test execution volume between 2014 and 20171. The bank also achieved hourly scheduled test runs on its core banking platform — a level of frequency uncommon in major Asian financial institutions2.
What it actually took: the platform reset behind the velocity
Engineering velocity at this scale is not a tool selection. The DBS public record makes clear that the velocity gains came on top of a multi-year platform reset:
- Cloud-native architecture. The bank moved aggressively into a virtual private cloud environment over the late 2010s, with the next-step roadmap targeting containerisation and hybrid public cloud for further productivity and scale3.
- Insourcing engineering. DBS reversed years of outsourcing-led IT, rebuilding internal engineering capability and establishing technology hubs across multiple countries.
- Platform-based operating model. Working with McKinsey, the bank reorganised around 33 internal platforms aligned to business segments and products, each led jointly by a business and a technology leader — what they called a "two-in-a-box" model4.
- Industrialised AI deployment. The same programme reduced end-to-end AI deployment time from 18 months to less than 5 months via an internal AI platform (ALAN), with the explicit goal of compressing this further toward a few weeks4.
What the data tells us about engineering as a competitive moat
The DBS case is one of a small handful of public benchmarks that quantify what platform-led DevOps discipline produces at incumbent-financial-institution scale. Three points stand out for any leader weighing a similar investment.
The cycle time gap is asymmetric. Industry research has separately documented that financial institutions actively pursuing DevOps transformation report meaningful improvements in deployment frequency, while institutions that do not are increasingly outpaced by digital-only entrants in the same regulatory environment5.
Testing velocity compounds. The published DBS Hong Kong figures — manual testing effort reduced from 20 person-days for a new function to 1 person-day for checking the test report, and functional test execution shrunk from five days to seconds — are the operational mechanism that makes a 10× release cadence safe in a regulated environment2.
The talent and culture decision is upstream of the technology. DBS's own public account makes clear that GANDALF would not have produced these outcomes without the deliberate insourcing of engineering, the internal "Digify" learning programme, and the leadership-level commitment to operating the bank as a technology company4.
What this means for incumbent financial institutions today
The DBS case is the clearest public answer to a question many bank, insurance, and asset-management boards are now asking: what does it actually take for a regulated incumbent to operate at digital-native engineering velocity? The answer is not a single tool, framework, or vendor. It is a platform-based operating model, an industrialised CI/CD and test-automation backbone, and an organisational commitment that treats engineering as core capability rather than as an outsourced cost line.
It is also clear that the timeline is multi-year, not multi-quarter. DBS's most-cited figures cover a four-to-five-year arc; the underlying operating-model shift began nearly a decade earlier and is, on the bank's own account, still continuing.
"Velocity at scale is not a tool selection. It is a platform-based operating model, an industrialised testing backbone, and the discipline to insource the capability that compounds."
Where NovasIQ comes in
NovasIQ's Technology Transformation, Systems Design & Integration, and Cybersecurity capabilities together address the full stack required for this kind of programme. Our DevSecOps work integrates security from commit-time, not as a late-stage gate. Our QA capability is built around the principle that test automation and continuous deployment are inseparable — you cannot have one safely at scale without the other. And our Managed Services pillar provides the operational uplift that prevents the "first-year miracle, second-year regression" pattern that defeats so many transformation programmes.
The capabilities that map to outcomes like these
- Technology Transformation — for the platform-based operating model and architectural reset
- Systems Design & Integration — for CI/CD pipelines, microservices, and the event-driven backbone
- Cybersecurity — for DevSecOps integration and regulatory-grade governance
- Quality Assurance — for the test-automation discipline that makes 10× release cadence safe
- Learning & Development — for the engineering capability uplift the transformation depends on