Third-Party Risk Management (TPRM)

A Guide for Financial Institutions

Why Do Procurement Decisions Carry Growing Regulatory Exposure?

Regulatory assessments focus on evidence linked to supplier decisions. Supervisory expectations span DORA, EBA guidelines, and PRA/FCA frameworks (depending on jurisdiction), requiring documented pre-contract risk assessment, entity-level registers of material arrangements, and visibility on concentration exposure at the point of supplier selection.

PS7/26 in the UK extends this requirement by introducing mandatory notification before entering or materially changing a material third-party arrangement, which means procurement decisions trigger formal regulatory processes that depend on complete and current risk documentation.

If a critical supplier becomes unavailable tomorrow, who in your institution holds documented accountability, and what evidence can be produced within 48 hours?

Regulatory Terms at a Glance

DORA (Digital Operational Resilience Act) – EU regulation, in force since January 2025, requiring financial entities to manage ICT and third-party risk, including oversight of critical technology providers.

EBA (European Banking Authority) – sets EU-wide guidelines on outsourcing arrangements, including how banks assess, contract with, and monitor outsourced service providers.

PRA (Prudential Regulation Authority) – the Bank of England body responsible for prudential supervision of UK banks, insurers, and major investment firms.

FCA (Financial Conduct Authority) – the UK’s conduct regulator for financial services firms.

PS7/26 (The PRA’s Policy Statement on Operational Incident and Third-Party Reporting) – published in March 2026, introduce mandatory notification before entering or materially changing a material third-party arrangement. Rules take effect 18 March 2027.

Where Does TPRM Create Exposure in Practice?

Many institutions operate with TPRM frameworks broadly similar to those they used three years ago, while supplier risk has continued to evolve. In 2025, more than 40% of cyber incidents reported to the FCA involved a third party*, while most institutions still check their critical suppliers once a year. The gaps tend to appear in four places:

Classification models do not consistently incorporate newer risk dimensions such as GenAI dependency, fourth-party exposure, and geographic concentration. This results in supplier assessments that do not fully capture how services are delivered today.

What this looks like in practice

A bank outsources document processing to a BPO center in a region that is later affected by geopolitical disruption. The center stops operating and the bank’s processing is interrupted. Geographic concentration had been recorded at onboarding but never monitored afterwards, leaving the bank unable to show the exposure had been reassessed.

Supplier financial stability, ownership structures, and risk indicators are often checked manually and on fixed cycles. Supplier data is frequently fragmented across systems and formats, making it difficult to maintain a current view of exposure. Tools have improved workflow coordination and reporting, but changes in supplier risk are still identified through periodic checks rather than continuous detection.

What this looks like in practice

A data analytics vendor supporting credit decisions is found to have ownership links to a restricted entity. The onboarding assessment conducted several years earlier did not include beneficial ownership checks. Detection is delayed, exit is forced, and credit processing is disrupted.

Procurement, IT, and risk functions assess different aspects of supplier exposure. Without shared classification and aligned decision rights, no single view exists at the point a commercial commitment is made.

What this looks like in practice

A SaaS provider is onboarded for a customer-facing application. Procurement, IT, and risk complete separate assessments using different definitions of criticality. Onboarding is delayed, and concentration exposure remains unresolved at go-live.

Each institution typically conducts its own assessment of the same supplier, generating cost and supplier fatigue without proportionate risk reduction. Shared due diligence ecosystems are gaining adoption, with thousands of suppliers already in scope and a recent extension into pooled supplier audits. This raises a question institutions increasingly need to answer: which assessments should remain proprietary, and which can be sourced from a community.

What this looks like in practice

A mid-sized supplier serving dozens of banks is asked to complete a separate due diligence questionnaire by each one, covering largely the same ground. The supplier deprioritizes the slowest requesters, responses arrive late and inconsistent, and several institutions end up relying on assessments that are months out of date.

*Source: FCA, “FCA confirms new incident and third party rules to bolster resilience”, 18 March 2026 (fca.org.uk)

How Is GenAI Changing TPRM on the Supplier Side?

GenAI is changing TPRM primarily by changing what institutions need to understand about their suppliers. TPRM teams themselves still rely on human judgement for material decisions. The sharper question is how suppliers use GenAI within the services they provide: how models are trained and updated, what data flows into them, and what changes in model behavior mean for the institution relying on the output.

This affects two parts of the supplier relationship:

Supplier selection requires assessment of how GenAI is embedded in the service, and how the supplier governs model updates and external dependencies.
Contracts need to address data usage, model dependency, and the rights available to the institution if model behavior changes.

Without these, GenAI-related exposure is difficult to identify through standard cybersecurity or financial due diligence.

Key Questions from Our Clients

How do we know whether our current TPRM framework is still fit for purpose?

Frameworks should support continuous supplier monitoring as a procurement capability, decoupled from the sourcing event itself, rather than limiting risk assessment to sourcing events or periodic reviews.

We have thousands of suppliers. Where do we start?

Criticality mapping identifies suppliers that would create operational or regulatory impact within defined timeframes, while AI-supported screening of supplier data supports prioritization across large portfolios.

How do we incorporate GenAI vendor risk into an existing framework?

Existing frameworks need two additions rather than a rebuild. Supplier due diligence should assess how a supplier uses GenAI within its service, how models are trained and updated, what data feeds into them, and how the supplier governs changes in model behavior. Contracts need explicit terms on data usage, model dependency, and the institution’s rights if model behavior changes materially.

Because GenAI risk sits inside existing supplier relationships rather than as a separate category, it is best incorporated into the same risk dimensions, monitoring cadence, and contractual review already used for other supplier risk factors, rather than being run as a parallel process.

What must remain local in a cross-border operating model?

Regulatory accountability remains with the licensed entity, while central coordination supports methodology and monitoring and local entities retain approval and reporting responsibilities.

How Inverto Applies TPRM within Procurement

Our work with financial institutions on TPRM covers three connected operating capabilities that establish continuous supplier risk management across the supplier lifecycle. Rather than treating TPRM as a sourcing activity, risk assessment becomes an ongoing AI-supported procurement capability that informs sourcing whenever commercial decisions are made. This enables faster, better-informed sourcing awards.

Designing the framework and governance

We work with risk, procurement, IT, and legal to define the risk dimensions, scoring methodology, and decision rights that determine how supplier risk is assessed and who is accountable at each stage. This includes alignment to regulatory expectations under DORA, PS7/26, and EBA guidelines, and ensures that the framework reflects current risk dimensions including GenAI dependency, fourth-party exposure, and concentration.
Segmenting suppliers and running TPRM as part of procurement’s daily business

We apply the framework across the supplier base to assign supplier tiering, map regional concentration and fourth-party dependencies, and continuously reassess supplier exposure as new information becomes available – decoupled from the sourcing event itself. GenAI-enabled monitoring supports daily screening of supplier signals across structured and unstructured data sources, allowing procurement teams to focus their attention on suppliers where material risk changes are detected. The output is an operating model that procurement, risk, and IT can run from a shared view, with the governance structure to support material decisions.
Fine-tuning risk assessment and mitigation in sourcing and contracting

With supplier risk already understood through continuous monitoring, well before a sourcing process starts, tenders and negotiations focus on calibrating the specific mitigations a deal requires, rather than beginning risk assessment from scratch. This includes audit rights, subcontractor disclosure, notification timelines, and exit provisions. Risk insights translate into commercial and contractual protections tailored to what the ongoing assessment has already flagged for that supplier. These protections are agreed during sourcing rather than retrofitted afterwards.