In 2025, the way we think about underwriting and risk assessment in the UK is changing. The shift has been slow, gradually gaining momentum due to the growth of embedded lending. For those working in credit risk, be it banks, lenders, brokers or aggregators, the focus has shifted to how underwriting and credit risk can be streamlined, automated and made to work at scale. 

Embedded architecture is forcing us to rethink risk assessment, how it is approached, measured, and who owns the decision. In this blog, we will explore how underwriting is evolving and what this transformation means specifically for the UK market.

From Point-in-Time Underwriting to Real-Time Flow-Based Underwriting
Historically, underwriting in the UK has followed a fairly structured path. The borrower initiates an application, the bank/funder collects documents, runs bureau data, applies credit scores, with some manual judgment, and post which, the loan is approved, and funds are disbursed. Risk assessment occurs only at one given point, usually at the time of origination. In rare cases, it is followed by periodic monitoring and then remedial measures ensue if things go wrong. 

With embedded lending, things work differently. The credit decision is embedded into the consumer or business’s journey. This can be at checkout, inside a SaaS portal or dashboard, or within a marketplace transaction. For example, recent studies show that embedded-lending platforms in the UK are offering credit decisions at the “moment of need” using high-quality, multifaceted platform-native data instead of archaic bank statements and bureau scores. 

How Does This Impact Underwriting and Risk Assessment? 

• Real-Time Data Streams: Instead of relying on lagged bureau scores, data feeds from Open Banking, Open Accounting, and alternate sources provide near-real-time inputs. Thus, underwriting becomes dramatically faster, powered by real-time data and artificial intelligence. 
• Contextual decisioning: Embedded lending inherently suggests the right funding solution, at the appropriate time of need, making lending contextual. That context becomes a risk driver. For example, a marketplace merchant seeking working capital might be approved not simply because of credit ratings, but because their transaction flow on the platform (orders, returns, supply-chain delays) shows a sudden rise in late shipments or delayed payments. The “moment‐of-need” model means underwriting doesn’t wait for a full application but triggers as part of an event or journey. 
• Speed and automation: Since the decision must happen inside a flow, human review is often impractical. Underwriting automation offers speed and reduces manual intervention, often providing near-instant decisions. Pulse’s Einstein aiDeal for instance, is capable of processing thousands of applications simultaneously, while auto-decisioning 95% of deals in under 45 seconds. That changes not just the process but risk oversight. 
• Embedded monitoring: Once a loan is originated, the monitoring doesn’t kick in only at scheduled intervals. Instead, it can be continuous, triggered by off-platform triggers such as a drop in revenue, a change in purchase behaviour, or churn. UK lenders are already using risk-assessment frameworks powered by AI and real-time data. 
• Thus, underwriting has transformed into flow-based underwriting with real-time risk monitoring. Risk models need to handle dynamically changing borrower contexts and event triggers, and incorporate alternative data. The shift also means risk assessment is more embedded in the tech stack. An excellent example would be Pulse’s Unified Lending Interface and its automated underwriting solution, Einstein aiDeal, which enables automated decisions for 95% of deals in under 45 seconds while leveraging alternative data sources like OA and OB. 

Embedded Data Ecosystems, Consent and Standardised Infrastructure

A key enabler of the underwriting evolution is the change in data architecture and infrastructure. In the UK, the interplay of Open Banking/Open Finance, regulatory frameworks, and fintech infrastructure is creating a new operating model for underwriting. 

Data Ecosystems and Consent
UK businesses are increasingly leveraging open banking data and other platform‐native signals such as marketplace order flows and SaaS usage metrics.
However, the consent layer is critical to underwriting. The borrower must authorise data sharing, and usually, the SaaS company mediates this. The underwriting model thus becomes more transparent as per UK regulations, including the Financial Conduct Authority (FCA) and data-protection regimes.  

Infrastructure and standardisation
While embedded lending offers speed and flexibility, one major bottleneck has been fragmented integrations and non-standard data flows. In 2025, Pulse pioneered the concept of a Unified Lending Interface (ULI), which is emerging: a standardised interoperability layer of APIs, data schemas, and event models for credit origination and servicing in the UK.  

Pulse ULI or similar frameworks matter because they enable: 

• Cross-platform data sharing, so a borrower’s exposures across multiple embedded lenders can be visible. 
• Consent and privacy control are embedded into the architecture. 
• Modular solutions, including Pulse’s Loan Origination System (LOS), Pulse’s Loan Management System (LMS), and Einstein aiDeal’s automated underwriting, which can be adapted and integrated into embedded flows. To learn more about Pulse ULI, contact us today.

Regulatory overlays and risk-governance
Embedded lending doesn’t occur in a regulation-free vacuum. The FCA’s increasing focus on affordability (notably in BNPL) and the broader regime handled by the Prudential Regulation Authority (PRA) means underwriting frameworks must embed compliance and monitoring. For example, embedded finance players must ensure that credit decisions and data-sharing remain “fair, affordable and appropriate. Pulse’s ULI has embedded security, protection and compliance in place, allowing for a seamless lending journey and customer experience. 

Thus, embedded-underwriting frameworks have to embed explainability, bias-mitigation, audit-trails and real-time data-governance as first-order features.

Where Underwriting and Risk Assessment Will Go from Here

Looking ahead from 2025, embedded lending underwriters and risk teams in the UK should gear up for: 

• Contextual-event underwriting: Expect more triggers tied to platform events (e.g., “merchant’s returns rate > X in last 14 days → credit line review”) rather than only new-loan origination. 
• Continuous model adaptation: Embedded flows mean data refresh cycles are compressed; risk teams must monitor model drift continuously rather than periodic recalibration. ML and AI governance will remain crucial. 
• Ecosystem-wide visibility: With ULI-style infrastructure, underwriting will shift from standalone to interconnected. Risk assessment will consider multi-platform exposures, cross-product stacking, and aggregated borrower risk. 
• Adaptive pricing and risk models: As improved risk visibility allows the inclusion of segments that often lack traditional credit history (thin-file, gig-workers, platform-merchants), pricing and risk models will evolve. Risk teams should collaborate with the product to balance accessibility and risk-adjusted returns. 

Conclusion 

Embedded lending is reshaping underwriting and risk assessment in the UK in a meaningful way. Rather than incremental change, we are seeing a shift from static, episodic underwriting to dynamic, integrated, ecosystem-aware risk frameworks. For UK practitioners, this means evolving data architecture, regulatory alignment, and model governance. Those who build underwriting engines which are flow-embedded, data-rich, model-driven, and ecosystem-centric will be best positioned. The risk function is no longer just Saturday-morning spreadsheet review; it’s an always-on, live layer across multiple platforms. Stakeholders can always choose to integrate with API-first solutions like Pulse’s ULI and leverage its solutions to digitise, automate, and streamline the entire lending journey, including post-disbursement servicing. This would allow users to focus on increasing volumes and scale while boosting profit without dealing with the massive upfront costs of self-building. Whether you choose to underwrite yourself or integrate solutions, underwriting has evolved substantially, and 2026 will see it grow further, especially with respect to embedded lending.