In 2025, modernisation continues to be one of the top agenda items for financial institutions. Innovative new use cases for technologies like AI and a wave of regulatory updates necessitate significant system and process modernisation, and with modernisation
comes the need for testing.
Many traditional banks want to include fintech solutions in their existing infrastructure, but there is a risk posed by using new solutions without being able to test them first. A safe way to do this is through sandboxes, which can test run new solutions,
preventing risk while encouraging innovation.
Sandboxes continue to be one of the most efficient ways for organisations to safely experiment, prototype and analyse proof-of-concepts without impacting live networks, systems or sensitive customer information. For example, after two successful pilots,
the
FCA launched a permanent Digital Sandbox in August 2023.
Through their controlled environment, digital sandboxes encourage innovation, mitigate testing risk, offer an opportunity for regulators and institutions to closely work together, and speed up time to market. All of these factors help foster a more competitive
financial landscape.
To find out about how organisations can effectively leverage digital sandboxes, and what’s needed from the digital sandbox model in 2025 and beyond, Finextra spoke to
Karan Jain, CEO at Naya One; Patrick Burgess, head of payments developer experience and connectivity at J.P. Morgan Payments;
Chintan Mehta, CIO, digital technology and innovation at Wells Fargo; and
Sam Dover, general manager - strategic partnerships and market development at Trustwise.ai.
Sandbox innovation in 2025
Over the last few years, digital sandboxes have quickly evolved from nice-to-have to a critical part of building and validating for many organisations. “Sandboxes used to feel like box-ticking exercises—limited data, limited scope,” commented Dover. “But
over the last few years, they've become real enablers of innovation. We’re now able to simulate real-world conditions, collaborate directly with regulators, and iterate fast—without compromising on safety or trust. That shift has been a game-changer for us
and the products we bring to market.”
Jain confirmed Dover’s observations. “Initially, their value was simple: give technology vendors a safe space to test against compliance rules. Today, we’re seeing a much broader evolution. Sandboxes are being used by major financial institutions. Not just
to meet regulatory obligations, but to drive innovation—testing AI, DLT, embedded finance, and open banking propositions in live-like environments. They’ve gone from risk-mitigation tools to launchpads for competitive advantage. We’re also seeing the rise
of thematic sandboxes focused on real-world problems such as ESG, scam prevention, digital identity, allowing for more focused experimentation. This model is no longer fringe. It’s becoming foundational.”
Datasphere Initiative research from January 2025 finds that today, there are 66 sandboxes related to data, AI or technology worldwide—59 of which are
national, while the remaining seven are global initiatives. Out of the 66, an impressive 31 are designed specifically to foster AI innovation, highlighting the crucial role of sandboxes in the transformation of financial services.
However, sandboxes are not just used to develop AI—Mehta explained how generative AI helps Wells Fargo optimise the sandbox itself. “We have developed a flexible, scalable, and secure platform that includes a sophisticated and state of the art generative
AI capability within the sandbox. This platform supports both cloud and on-prem environments and is provider (AI model, for example) agnostic. This allows us to tune and optimise every layer of the stack. In the case of GenAI, think of it as supervised fine
tuning or distillation or quantisation,” he stated.
“Currently, we evaluate new ideas rigorously before moving the most promising ones into pilot use cases and then development. Efforts to accelerate delivery through improved operating models and processes have led to a 25% speed improvement in delivering
solutions like the branch procedure virtual assistant.“
Today’s use cases for digital sandboxes have expanded far beyond the initial scenarios of compliance, explained Jain. ”Now, clients are just as focused on innovation, particularly around GenAI, data strategies, and embedded finance.
“We’re seeing financial institutions using sandboxes to test AI-powered assistants, synthetic data for fraud models, and automated document processing at scale. In parallel, they’re trialling new payment flows, alternative lending platforms, and hyper-personalised
investment tools. While compliance remains critical, the balance has tipped. Sandboxes are now equally about product-market fit, user experience, and scaling new revenue models.”
However, it’s not only the ability to safely test that make sandboxes an ideal facilitator for innovation. The speed of testing is an additional advantage that sets them apart and is critical in today’s rapidly evolving technological landscape.
“Digital sandboxes can be spun up quickly using modern DevOps and CI/CD practices with test data to quickly prototype, test, and deliver new capabilities,” explained Burgess. “In comparison, traditional testing processes can be cumbersome to setup, especially
in the financial sector where many transactions go through mainframes that don’t lend themselves to spinning up a new instance quickly.”
As incumbent banks face increasing competition from digital-first neobanks, the speed with which they can bring new products and services to market is a crucial differentiator to stay competitive. Sandboxes allow them to do just that, if deployed effectively.
How to go from experimental attempts to opportunistic assembly
No innovation comes without its challenges, and a digital sandbox is no exception. One of them is the tightrope balance between the time necessary for experimentation and production: “It’s not a concern per se, but a typical aspect to keep a close eye on
is the ‘hardening’ time,” explained Mehta.
“Meaning if experiments can be done very quickly, but they take a long time to make it production ready, then it’s a challenge to the project’s ROI. Alternatively, if it’s too hard to do an experiment because you have the same constraints as the production
grade set-up, then it becomes too rigid for trying new things. The end state is to find a right ‘goldilocks’ zone where it’s balanced.”
Secondly, digital sandboxes may not have all the scenario and business logic of an end-to-end transaction lifecycle, explained Burgess. “Though digital sandboxes can provide full featured capabilities for simple APIs, especially with standardisation and
automation, the financial industry has a lot of business logic embedded in legacy infrastructure, such as monolithic infrastructure and mainframes. As such, it is not easy to provide all the functionality in a digital sandbox; in this case a hybrid model will
need to be considered.”
As with most infrastructure, the key lies in planning, and deploying for the appropriate use cases for each financial institution. “The opportunity is clear: digital sandboxes give financial institutions a fast, low-risk way to explore new tech, validate
vendors, and get ahead of regulatory curveballs. They create a space for structured experimentation, and when done right, drastically reduce time-to-market and procurement friction,” emphasised Jain.
“Challenges remain. Data privacy, legacy system integration, and internal alignment can slow things down. Translating test results into production still requires strong governance, technical maturity, and clear ownership. And unless the sandbox environment
stays tightly aligned with live systems, insights risk becoming outdated fast. Done well, though, the payoff is huge: faster learning, smarter adoption, and better innovation ROI.”
Data is key aspect of a successful sandbox deployment. “Digital sandboxes, if done incorrectly, can quickly get out of sync with the actual behaviour of the API,” Burgess added. “To maintain the data and functionality, this requires constant upkeep, and
because it is not the actual code itself, but rather a simulator or mock, the accuracy of the digital sandbox can drift.”
Data privacy is crucial, which institutions can achieve either by using synthetic or anonymised data, as Wells Fargo does. “Our digital sandboxes bring product and engineering groups together to develop Proof of Concepts (PoCs) using synthetic data through
our open banking API gateway platform,” Mehta stated. “This approach allows for the mechanisms we need to iterate and experiment.”
What does the future of the digital sandbox look like?
Sandbox adoption has been scaling fast and is showing no signs of slowing down. “We expect to see a shift from one-off experiments to continuous, integrated use,” Jain predicted. “Sandboxes will be always-on environments, embedded into the way banks and
technology vendors validate new tech, engage with regulators, and accelerate go-to-market. Institutions are realising that innovation can’t happen in isolation—it needs structure, speed, and compliance built in from day one.
Burgess expects the reach of sandboxes to expand: “More traditional organisations will adopt digital sandboxes such as clearing houses, which will increase innovation in the traditional banking industry and provide ways for fintechs to easily connect, test
out, and develop new financial products and services.”
“Regulators are getting more hands-on too,” added Jain. “Expect more thematic and cross-border sandboxes focused on areas like ESG, responsible AI, and financial inclusion. These aren’t just policy pilots, they’re becoming tools for shaping smarter regulation
in real time.”
Why sandboxes should be central to your strategy
In 2025 and beyond, digital sandboxes are one of the most effective ways for financial institutions to brave the regulatory storm, efficiently modernise, and ensure continued competitiveness.
“Digital sandboxes are pivotal in driving innovation in the sector. They provide a lower cost way to try out new use cases, quickly iterate on design of an API or application and discover new capabilities,” Burgess concluded.
As the pace of change accelerates, digital sandboxes are becoming central to how the financial sector experiments, collaborates, and scales new ideas.
“From where we sit, digital sandboxes are where real innovation starts,” added Dover. “At Trustwise.ai, we test ambitious ideas without waiting months for approvals or risking unintended consequences. Especially in regulated industries, where the cost of
getting it wrong is high, sandboxes give leading-edge companies the confidence to explore and scale, both internally and with customers.”
“They’re no longer just testbeds. They’re how the sector moves forward, together,” Jain closed. “We’re seeing sandboxes directly impact outcomes—from faster go-to-market timelines to higher funding rates for participating technology vendors. In 2025, digital
sandboxes will be where roadmaps are shaped, not just tested. They’ll help institutions make faster decisions, de-risk adoption, and stay ahead of the curve in a market that’s moving faster than ever.”