2025 was the year economic change finally outgrew the tools meant to track it. From a labour market that left traditional models behind to the rise of automated fraud, ‘good enough’ is no longer sufficient, writes Artem Lalaiants, CEO and co-founder of RiskSeal.
It has been a year where the pace of economic change finally outpaced the tools meant to track it. Many credit providers reached a point where balancing outdated risk routines with growth targets simply wasn’t sustainable anymore.
Here are the critical lessons learned from running an alternative data platform in 2025, and the trends that will determine who sinks or swims in the year ahead.
The trends that broke the scoring system
2025 exposed the labour trends that traditional models couldn’t predict. The global labour market continues to move away from traditional full-time roles—a shift that accelerated after the pandemic and never slowed down.
This coincided with a wave of white-collar layoffs and slower hiring, pushing many professionals toward freelance, contract, or project-based work. These income patterns are harder for traditional financial systems to recognise and assess.
What we have seen is millions of people suddenly showing up as ‘unbanked’ or ‘underbanked’ on paper. Not because they chose to step outside the system, but because the system just doesn’t match how they earn and spend anymore.
Gen Z redefines ‘credit history’
Gen Z makes the gap in traditional scoring impossible to ignore. They possess real income but often lack history because they rent and freelance, skipping the milestones credit bureaus typically rely on.
The missing insight lies in their digital lives. Their extensive online footprints provide the behavioural data needed to reveal how they actually live and earn.
The burden isn’t on the borrower to mimic the past, but on the lender to understand the present. We don’t need to wait for history to build; we need to change how we capture it. The industry finally has all the tools to flip the scoring model safely and compliantly.
Fraud as an automated industry
Digital channels now dominate fraud. Approximately 80 per cent of scams start online, and criminals move easily across apps, devices, and platforms. What used to be the exception is now the norm; we know that one in 20 verification attempts in digital banking is now fraudulent.
These attacks are largely automated and often reused across multiple institutions. The fraud fintechs face today bears almost no resemblance to what they fought even five years ago. It is an automated industry now. We are realising that old defences are simply unable to distinguish between a real human and a synthetic profile perfected by AI algorithms.
For a long time, there was caution around alternative data, as if it were too new to trust. That fear is gone. Fintechs are done theorising; they are rolling up their sleeves to bake this data right into the core of their decision-making. It is not a side project anymore—it is becoming the standard.
Three big changes coming to credit risk in 2026
The pressure on both credit organisations and alternative data providers is growing, and 2026 will test how fast the industry can adapt.
1. Agentic AI moves from pilot to production
AI is becoming a core part of underwriting, fraud detection, compliance, and portfolio optimisation. By 2026, truly agentic systems will not just make recommendations; they will take action and run processes end-to-end. However, granting machines this level of decision-making power creates a new dynamic for risk teams.
Agentic AI creates a paradox: the smarter the system, the more sensitive data it demands. That includes behavioural and contextual signals that risk teams never had to manage at this scale. The only way forward is to teach these agents to be powerful, but also privacy-aware, so innovation doesn’t come at the cost of control.
2. Regulatory oversight catches up
Regulators are reacting to AI’s reach with stricter governance, marking 2026 as the year oversight tightens. Meanwhile, a lower-rate environment is squeezing fintech margins.
To sustain growth, institutions are now prioritising three key levers: automation for efficiency, deeper risk analytics, and diversifying into fee-based streams like private credit and embedded services.
3. Embedded finance expands the fraud surface
Embedded finance and Banking-as-a-Service (BaaS) are surging as non-financial platforms integrate payments and lending directly into customer journeys. At the same time, real-time payments and digital wallets are becoming standard infrastructure.
This shift offers seamless speed for users, but it also creates a high-velocity environment that fraudsters are eager to exploit. Embedded finance is forcing non-lenders to become risk experts overnight. They need alternative data to handle onboarding and fraud, but there is a danger of falling into the ‘data hoarding’ trap. Success is about cutting through the noise to interpret only the signals that actually drive ROI.
Smarter data use
In 2026, the winners will be the institutions that treat data as a unified ecosystem rather than viewing risk, fraud, and growth as separate silos. Leaders must look beyond the immediate ROI of a single tool and focus on building an infrastructure flexible enough to handle the unknown.
The technology is ready. The real competitive advantage now lies in the courage to trust it and deploy it at scale.