Innovation at Scale: a Roadmap for the Financial Industry

The financial markets have been in a swindling state since the start of this decade. The prolonged uncertainty is a bountiful breeding ground for short-term thinking: The crypto-craze, decentralized finance, and, lately, generative AI. Some of the hype has been misdirected. Especially when R&D funding was abundant.

As the tide has changed and ROI reports got a sour response from the boardrooms, financial leaders must once again rethink technology adoption and product development.

Profitability is still lagging. Operational efficiencies have hardly improved, and customers are marginally happier.

Over half of consumers believe their bank isn’t interested in earning them money, and 27% say their bank offers unsuitable financial products.

2023 Digital Banking Experience Report

Many leaders lack a vision for spending on the right initiatives despite having the funds and technology.

In this post, we’d like to share our field notes on the 8 major themes in the financial sector and advice on using them to build a realistic innovation roadmap.

The AI Revolution in Finance Reaches a Critical Juncture

Revolutions may seem like blips on a historical timeline, but they don’t happen overnight. There’s a boiling build-up before the old epoch moves.

ChatGPT topped the headlines in late 2023, but the Generative Pre-trained Transformer (GPT) model was introduced in an OpenAI paper in 2018. Artificial neural networks (ANNs), used to train computer algorithms for cognitive tasks, have been around since 1956.

In 2018, Gartner placed deep neural networks at the top of its annual Hype Cycle, suggesting mainstream adoption within two to five years.

Software Engineering

Want to discuss software innovations for your financial business?

Contact our FinTech and Banking IT team

As of 2024, only a few Deep Tech companies have productized deep learning, such as OpenAI and Google, which used DL techniques to speed up language understanding for their transformer models.

In finance, machine learning algorithms have been used for decades for fraud detection, robotic process automation of back-office tasks, and automated customer support chatbots. Is this why surveys show that 99% of financial service companies have deployed or used AI?

This naturally begs the question: What constitutes AI adoption?

Does piloting a machine learning model for document analysis in one department count as “using AI”? For investor approval, yes. The same goes for letting marketing employees use a content marketing platform with Gen AI features (e.g., text or image generation).

So far, most successful AI implementations have been confined to a limited number of use cases in areas like operations, compliance, sales, marketing, and customer support.

Charting the tomorrow: A roadmap for financial product innovation

Source: State of AI in Financial Services 2024 by NVIDIA

Adopting ‘plug and play’ AI tools and experimenting with custom machine learning algorithms (whether they make it to production or not) are important pillars of the AI revolution.

But to turn the page and become AI-driven, financial businesses need to confront an old foe: data silos.

Every new technology, added to the legacy core systems over the years, creates data silos — customer, transactional, or operational data that isn’t easily accessible due to its storage.

Mortgage brokers can’t access customer transactional records for underwriting. Compliance lacks visibility into fraud volumes. Each department operates its own ‘data island’. This disparate mess undermines productivity, profitability, and propensity for AI adoption.

95% of banks say legacy systems and outdated core banking modules inhibit efforts to optimize data- and customer-centric growth strategies.


Every machine learning or deep learning use case, including risk, prospecting, marketing, customer experience, and fraud prevention, requires substantial historical and real-time data.

JP Morgan, ranking #1 in AI maturity on the Evident AI Index, has a 200-person AI research group for strategizing new AI applications, according to its Investor Report.… and over 1,000 are involved in data management for AI systems.

The NVIDIA report also notes that ‘data analytics’ and ‘data processing’ are the main AI workloads for financial institutions. And the biggest adoption challenge for AI? Data silos: privacy, sovereignty, and disparate data storage locations.

Without strong data engineering practices and a proper data governance framework, financial companies can’t advance beyond basic use cases of rule-based data processing, text summarization, or document classification. They’ll keep hitting the regulatory barrier.

AI adoption creates new compliance and cyber-security risks. Spilling sensitive corporate data results in regulatory fines. Poorly-tuned automated decision-making models (and lack of human oversight) lead to bigger problems.

France’s financial regulator fined Credit Agricole €1.5 million for AML violations due to its automated risk engines supporting an incomplete set of scenarios that failed to detect certain atypical transactions. A German regulator issued a €300K fine to a bank for lacking transparency in automated credit card application rejections. Charles Schwab paid $187 million to the SEC for providing misleading information through its robo-advisory service.

If you build an AI/ML product in data silos, you’re going to get yourself in trouble. And double-trouble if you use ‘off-the-shelf’ Gen AI tools without understanding how they work.

As Matt Mollison, chief data scientist, Branch International, said during a Money 20/20 panel:

“There are two main things that push me away from using Chat GPT for underwriting, and that’s that we’re not in control of the data that’s been trained on. And we’re not in control of the model itself. We don’t know all the parameters of land, the model, the architecture; you want to know the architecture, the parameters, and be able to explain the predictions that it’s making”.

To understand, get your data management in order. Look into your data silos and work on data labeling, classification, and lineage management. Only when financial companies have full control over key metadata and data streams can they run AI-powered operations without risking compliance, tech-related, or cybersecurity backfires.

Real-time Payments Require Real-time Action

Bank transfers take three days to clear in many developed countries, about two-thirds the time of the Titanic’s voyage. Let that sink in for a sec.

Long clearance times aren’t just frustrating. They’re also undermining the financial systems’ performance. The European Commission estimates that €187bn is locked daily in the ‘being transferred state’ — a grand sum that could be put to better use if being liquid.

Regulators aim to fix that. The European Commission Instant Payment regulation, approved in early 2024, mandates banks to create a real-time payment infrastructure for processing all transfers within 10 seconds, anytime day or night, and without extra fees.

In the US, the Feds launched an instant payment service, FedNow®, in mid-July 2023. FedNow allows banks and credit institutions to eliminate the dreadful several-day lag for account-to-account (A2A) transfers. At the start of 2024, only 400 institutions had signed up, with many only opting to receive rather than send real-time payments.

That’s a shame because A2A transactions could account for about $200 billion in consumer-to-business transactions by 2026 in North America — and even more globally.

Card-issuing banks that take a timid approach to payment innovation can lose out on 4.6% of total global card and online payments revenues, or $89 billion, in the next three years.


In many cases, banks are reluctant to get on real-time rails for two reasons.

Higher fraud rates are an unjustified concern. Real-time payments are secure because each transaction is authenticated by a consumer’s online banking credentials. With proper fraud-monitoring systems, multi-factor authentication, confirmation of payee, and security guardrails, make instant payments as secure as traditional transfers.

VocaLink by Mastercard, the technical provider of Faster Payments infrastructure in the UK, created a Mule Insights Tactical Solution (MITS), powered by machine learning and advanced analytics, to prevent authorized push payment (APP) fraud.  The Bank of Thailand created security guidelines for instant payments, requiring FIs to have a system for suspicious transaction detection, digital ID authentication, and advice to use biometrics to authenticate users for new bank accounts and transactions above a regulatory threshold.

Anti-fraud techniques for instant payments

Beyond the hype: What the financial sector needs to accelerate digital revenues

Source: World Bank

The second more сompelling reason is legacy system architectures. Core systems from the 1960s were not designed for high transaction volumes. Newer monolith systems only support batch processing and are difficult to integrate with external data sources or new payment processors.

In such cases, Sigma recommends building a cloud-native orchestration layer around the core payment systems to enable data exchanges and integration with shared services like fraud detection or AML services. This system re-routes payment data from the legacy core to a new cloud-based infrastructure with globally distributed processing capabilities, in line with industry regulations.

By moving payments to the cloud, FIs get higher service scalability, availability, and eliminate data silos. This enables real-time access to transactional data for analytics and machine learning systems for KYC, AML, and customer-centric innovation.

By understanding your customers’ spending habits, you can suggest personalized financial products or create new revenue models. With a cloud architecture, you can embed partner offers using APIs for new services like buy-now-pay-later, instant credit lines, or tax loss harvesting. You can also populate your platform with new partner offers — insurance, investing, or lending products, which brings us to the next point.

Open Banking and Embedded Finance Open Opportunities Galore

In 2022, Open Banking generated over £7.2bn in revenue opportunities in the UK. Following Europe and Australia, the US issued an Open Banking proposal in late 2023, allowing customers to share their checking, credit, prepaid, and digital wallet accounts.

Open Banking and BaaS, as its continuation, reduces customer acquisition costs, improves retention and profitability, offers new revenue diversification models, and ease innovation.

Effectively, banks can choose to become a marketplace player, curating a mix of internal and partner-supplied financial products or a platform player — supplying its products and financial infrastructure to non-financial partners.

Starling Bank offers a product marketplace for customers to compare insurance, get cashback from its partner Tail, invest with Wealthify, or improve their credit score with CreditLadder — all within the app. The bank’s average revenue per customer doubled between 2022 and 2023, and revenue increased by 109%.

The number of US banks offering BaaS partnerships increased from 100 in 2019 to 500 by the end of 2022. Green Dot Corporation has Apple, Amazon, and Uber as BaaS clients, and continues to sign up new FinTechs, ranging from on-demand pay solutions and small business lenders to wealth management platforms.

FinTech partnerships are just the beginning. Sigma team has identified embedded payment and lending opportunities in industries like automotive, gaming, retail, and aviation.

Connected vehicles now come with a full-feature OS and plenty of apps, where consumers can spend money on marketing, ordering drive-through, or premium infotainment. Automakers like Stellantis and GM aim to generate over $20 billion from in-car commerce and subscription services by the end of this decade — and they’ll need financial infrastructure providers.

OEMs and manufacturers can benefit from direct to consumer (D2C) sales, enabled by your bank. For instance, Volvo Penta now sells over 150K parts and accessories in 15 markets via an ecommerce portal, built by Sigma Software.

Partnerships with airlines are a great way for customer acquisition. A 12-hour flight gives flyers time to study the airline credit card deal on their infotainment screen and complete the application (and even get approval with data-driven underwriting). We helped a financial client launch a digital mortgage processing platform, offering 24-hour approvals.

Likewise, banks can capture transactional revenues by offering travelers cab bookings, car rentals, dining reservations, or grocery restocking to avoid returning to a sad, empty fridge.

By 2030, 74% of global digital consumer payments will take place on platforms owned by non-financial institutions.

International Data Corporation

Banks can either lose this channel to payment services providers or FinTechs or start building their platform strategy today.

By combining cloud-native microservices architecture with an API strategy and strong data governance processes, FIs can benefit from the ‘economy of scale’ and grow vertically and horizontally.

You can save up to 12 months in product development by integrating new financial products or in-demand features from partners. Sigma team developed a new digital banking system within 9 months with standard MVP features, including a core banking system, using BaaS components. By implementing a scalable, high-load architecture, the bank can quickly add new features and launch embedded offerings without risking performance issues.

Time to Redraw Customer Journeys for ‘Lifestyle Banking’

Brett King got the industry talking about Banking 3.0 and then Banking 4.0. But the truth is: Consumers no longer want a bank per se. They want a financial partner — and it doesn’t have to be a traditional financial institution.

Two in five young US consumers believe the bank of the future will be a tech company, and 65% expect their primary bank to be fully online by 2027. Even today, consumers have plenty of companies to trust with their money, from Apple and CashApp to Target and Walmart. According to PYMNTS, 68% of consumers avoid using one app for all their finances.

In the B2B world, merchants have many options for financial services, from their accounting provider like Quickbooks, a POS provider like Square, or a niche FinTech provider, like the emerging cannabis banks.

Unless heritage financial institutions adapt to new preferences, they risk losing oodles of customers to non-financial players.

Banking agents spend only 9% of their time on customer interactions, a Capgemini report found. Globally, banks are losing 20% of customers due to poor experience, according to a 10X Banking survey.

Customer preferences have changed, but engagement mechanisms have stayed the same, with many banks lagging in omnichannel service delivery, product UX, and relationship building. Almost 65% of banking leaders admit that slow digital transformation has resulted in them missing out on winning new customers.

Software Engineering
Looking for an expert banking software development team to take your banking solution to a new level?

Contact our Banking Center of Excellence

Comprehensive customer intelligence is crucial for better engagement strategies and product development. If your current infrastructure can’t provide this, focus on backend transformation first, then the front end.

Unblocking data access leads to informed decision-making at every customer lifecycle stage. UAE bank Mashreq uses historical data to predict campaign ROI and optimize customer acquisition costs.

Analytics helps identify customer behavior at scale without relying heavily on the branch and 1:1 convos. A client who signed in five years ago while still in college may now earn a six-figure salary while your bank still markets them low-interest savings accounts. Predictive customer analytics tools and next-best-action systems can turn you into a lifestyle partner, delighting them with the right offer at the right time.

The roadmap to digital excellence: How to level-set tech investments with profitability

Source: McKinsey

Customer-driven Product Development Sets the Path to Profitability

With plunging profits and growing loan losses, traditional banks are in belt-tightening mode, with 15% planning to cut their digital transformation budgets. FinTech firms also face a tougher reality with limited VC funds, as funding volumes dropped by 70% YoY between Q2 2022 and Q2 2023.

When money was abundant, building innovative (or just copycat) financial products was a matter of tech experience. Everyone raced to pack more financial products into one superapp to grab a bigger market share from incumbents.

FIs need to improve their product development roadmaps. The best way to do so is to put the ‘customer’ back into the center.

When the first FinTech companies rose to fame, they focused on doing one thing well: Listening to the market’s demands for no-fee bank accounts, fast P2P payments, and easy online payment processing tools. What changed since then is that many product owners stopped thinking about customer needs and started obsessing over hyped market trends.

But 67% of consumers aren’t interested in using an AI assistant for managing their account. They want a simpler digital account opening process, real-time alerts on account activity, and the ability to aggregate financial data from multiple accounts.

High inflation discourages customers from keeping money in regular accounts. Launching a high-yield savings account (HYSA) or pitching a retail investment product can be the right move. Investment market calamities prompt HNW individuals to explore alternative investment classes. Offer access to tokenized assets like fractional ownership in real estate or art to profit from portfolio management and safekeeping fees.

As TS Anil, CEO of Monzo, rightfully observed:

“The incumbent banking industry is disproportionately built-off a business model that leverages one or more of the three things: Customer ignorance, customer vulnerability, and customer inertia”.

Monzo challenged the status quo by focusing on “making money when our customers do things that are good for them.” As of 2023, Monzo services 7.5 million customers, with 1.6 million acquired in the last year (when the market sucked). Revenue grew by 2.3X in FY2023, with the bank reaching profitability. Card spending increased by 28%, deposits by 34%, and the bank consistently racks high praise for its CX, banking app usability, and new product features.

Anything that goes on the roadmap must be customer-centric. Get to know your customer (beyond KYC) before investing money in new shiny features and technologies.

Conduct customer workshops and UX analysis (areas where Sigma will be happy to chime in) to shape your product roadmap. Build feedback loops to bring ideas from users’ lips straight to product owners’ ears and then snappy developers’ finders. Allocate investments towards ideas that can generate long-term value and profitability.

Digital Excellence Requires Process Excellence

During rapid digital growth, many FIs doubled their customer bases and workforce. The problem? Many didn’t scale internal processes to support jumbo-sized teams.

Business models are less stable. To survive market flux, you need to constantly adapt and make the right choices in planning, execution, and optimization at twice the speed.

When it comes to execution, few companies succeed. McKinsey found that only 10% of companies undergoing agile transformation report it as highly successful. Old habits die hard, especially when backed by legacy software, bloated org charts, and corporate inertia. But consumers are losing tolerance for all of the above, and leaders recognize that.

“I see a threat from fintech firms trying to dig into our business model or in that other insurers have long since jumped on the agile bandwagon and designed their products much faster and more flexibly”.

Participant in a University of Innsbruck study of Agile transformations

Many lack a method for transitioning from a hierarchical, operationally siloed organization to a customer-oriented, digitally-enabled leader. Why? Agile transformations require knowledge outside financial leaders’ purview — cloud-native system design, engineering process excellence, feature-driven development, and lean product development.

What sets Agile financial companies apart is the ability to:

  • Produce a product vision based on customer feedback and senior leaders’ strategic vision.
  • Cross-functional teams working in parallel on different items from the product development roadmap using modern technologies.

Agile companies use collective intelligence to source product ideas, solve inefficiencies, and test new revenue channels instead of relying on a centralized ‘innovation’ team. Using Agile practices, JPMorgan Chase reduced the standard development lifecycle for new mobile app features from over a year to under six months.

“The key to agile transformation that leads to product transformation is building cross-functional teams with clear accountability and rapid decision-making autonomy. In contrast to traditional work models that isolate each team to tackle challenges in silos, the agile process connects teams,” he says, “so our engineers always know the optimal ways—and reasons—to build, test, and manage solution-oriented software.”

Rohan Amin, chief product officer at Chase

Danske Bank, another Agile adopter, gained operating efficiencies by merging redundant projects, restructuring core teams across Business and IT to become product owners, and implementing better DevOps processes. Now, the bank can ship MVPs in under 4 months.

To gain agility, companies need a governance approach, allowing rapid resource redeployment to high-yield activities. This enables quick resource mobilization for technological or market opportunities. Bringing in external advisors experienced in Agile processes to drive cultural transformations. Sigma helped TEBIN, a fast-growing scale-up, improve project management, change management, and delivery processes, resulting in a 30% increase in operational efficiencies.

Align Your Infrastructure with New Regulations

In a tight market, FIs face tough regulatory scrutiny (understandable after the Silicon Valley Bank bust and Credit Suisse bankruptcy). FinTech companies are also under the magnifying glass. In 2023, 93% said compliance is more challenging, and 37% already coughed up over $500K in fines.

Regulators demand more than just ‘checkbox’ compliance amid heightened risk. They want FIs to show better risk modeling, tighter underwriting, and improved credit portfolio stress testing.

New regulations are rolling in 2024, many related to technology use.

PSD3 directive contains new measures to mitigate payment fraud and protect customer rights. NIS 2 Directive on cybersecurity also concerns banks and financial market infrastructure companies. The Digital Operational Resilience Act (DORA), effective from 17 January 2025, imposes new provisions for ICT risk management, information sharing, and digital operational resilience testing.

These regulations will require changes to existing IT infrastructure and ongoing security monitoring and pen test reporting. When in-house teams can’t meet regulatory requirements, involve third-party consultants. Sigma offers DORA compliance audits and end-to-end support with policies, procedures, technical controls implementation, and staff training. We also advise on compliance, including GDPR, NIS 2, PSDS, and OWASP, among others.

One Last Crypto-penny for Your Thought: the Bond with Blockchain

While AI may be the ‘du-jour’ hyped technology, banks aren’t letting go of another media darling: blockchain.

In 2023, Goldman Sachs, BNP Paribas, Deloitte, and 30 other firms launched Canton Network — a global blockchain network of tokenized assets. The pilot proved the feasibility of using decentralized technology for financial transactions. JPMorgan Chase and Citibank are trying to persuade Wall Street to use blockchain to reduce settlement times and costs. In the meantime, over 100 central banks are conducting feasibility studies for central bank digital currencies (CBDCs).

On the retail side, 75% of US consumers want to buy crypto-currencies from their bank. Challengers like Revolut, Chase, Ally Bank, Chime, and Monzo offer this, as well as incumbents such as Goldman Sachs, Barclays, and JP Morgan Chase. The U.S. Securities and Exchange Commission (SEC) approved a dozen ETFs tracking the world’s largest cryptocurrency at the start of the year. So, trade volumes will likely surge in 2024. the world’s largest cryptocurrency. So, the trade volumes will likely surge in 2024.

Regulators are ready to bond with the blockchain and include crypto coins in the financial system. The Markets in Crypto Assets Regulation (MiCA) proposes a legal framework for EU Markets, tracking crypto-asset transactions and increasing investor protection. The UK plans to release a new regulatory framework for cryptocurrencies and stablecoins by July 2024. In the US, a pro-blockchain bill was cleared by Congress at the end of 2023.

Globally, 52% of consumers believe cryptocurrencies will be an “important asset class and method of payment transactions” in the future. Regulators seem to share this sentiment, suggesting a new quest for blockchain use cases in finance is likely.

Charting the Right Course

Building financial products is difficult—especially today, when the markets are in flux and consumer behaviors change quicker than a hiccup. Rather than lamenting the constraints, a better use of energy is to create an efficient way to capture market sentiment and transcribe it as jobs to be done by your customers.

Knowing why your customers (or internal users) make the decisions they do and what they want to accomplish will help you find the right areas for innovation — and Sigma Software can help you find the right technologies to deliver these. Contact us to learn more about our technology implementation and consulting services.

Share article: