Untitled

Table of Contents

Finastra improves banking efficiency with AI

  • Finastra introduced OperatorAssist, an AI-powered feature to speed up the management of incidents and exceptions in payments.
  • The tool integrates into the company’s payments hub and aims to reduce errors, resolution times, and operating costs.
  • Initial results cited by the company point to efficiency improvements of more than 20% and a 20–30% reduction in manual investigation time. These metrics correspond to early estimates communicated by Finastra and may vary depending on the volume of exceptions, internal processes, and the degree of adoption.
  • It is available as an add-on for Global PAYplus and Payments To Go users, on a cloud-native and ISO 20022-native basis.
Mentioned metric What it means in payments operations Status of the figure (as published) Conditions that typically affect the outcome
Efficiency >20% Overall productivity gain in handling exceptions (less time per case / more cases closed per shift) Early result communicated by Finastra Exception mix (simple vs. complex), data quality, discipline in using the guided workflow
20–30% less manual investigation Less “search” time (reconstructing history, reviewing fields, identifying the likely cause) Measurable impact reported by Finastra Level of process standardization, availability of traces, ISO 20022 consistency
1.5+ hours/day saved per user Time freed up for higher-value tasks (prioritization, communication with teams, quality control) Impact claim reported by Finastra Daily volume of exceptions per operator, adoption curve, queue and role configuration

Introduction to OperatorAssist and its impact on payments handling

Finastra, a global provider of software for financial services, announced the launch of OperatorAssist, an artificial intelligence-powered solution designed to help

banks to manage the payment lifecycle with less operational friction. The focus is on one of the most costly points in the back office: payment errors and exception handling, which often require manual investigation, coordination between teams, and multiple iterations until a case is closed. In practice, an “exception” usually refers to any payment that cannot be completed automatically and needs operational intervention to diagnose, repair, or route it to resolution.

According to the company, OperatorAssist is integrated into the interface of its payment hub to speed up incident resolution and reduce costs. In a context where customers expect traceability and speed—and banks seek to increase straight-through processing (STP)—the promise is clear: less time “stuck” on complex cases and more capacity to operate at scale.

Payment exception management
– Where it fits: OperatorAssist targets the operational segment of the payment cycle where a payment doesn’t “go through” automatically and drops into an investigation queue (exception).
– What an exception is (in practice): a payment with incomplete/inconsistent data, failed validations, routing rules that don’t apply, or repair requirements before sending/confirming it.
– Why it affects STP: each exception reduces the straight-through processing percentage and adds cycle time (and cost) because it requires human intervention, coordination, and traceability.
– Time context: the launch was publicly announced on March 5, 2026, so the shared metrics are presented as early results/estimates.

Key features of the OperatorAssist solution

Automation of analysis and repair recommendations

OperatorAssist applies AI to analyze incidents, identify possible causes, and recommend repair actions. In practice, this aims to replace part of the repetitive work of reviewing data, reconstructing the payment history, and deciding the next step, guiding the operator through the resolution process.

Barry Rodrigues, EVP of Payments at Finastra, framed the launch as a meaningful change in day-to-day operations by combining AI with a cloud-native and ISO 20022 platform, with the goal of “eliminating friction” and enabling “faster and smarter” ways to resolve issues.

Exceptions Flow: Before and After
What the “before vs after” flow typically looks like in

an exception
– Before (more manual)
1) Detection: the payment falls into a queue due to error/validation.
2) Analysis: the operator looks up history, reviews fields, consults rules and evidence.
3) Repair: corrects data / reroutes / requests information.
4) Closure: documents, reports, and confirms resolution.
– With assistance (more guided)
1) Detection: the case appears with context within the hub.
2) Assisted analysis: the AI suggests a likely cause and highlights relevant fields.
3) Recommended repair: proposes actions and steps to execute/validate.
4) Closure with traceability: guides the logging of what was done and reduces rework.
Useful operational checkpoints: if the recommendation does not match the case evidence, if key data is missing (e.g., incomplete ISO 20022 fields), or if the case is “edge” (infrequent), it is advisable to escalate or force a human review before executing repairs.

Improvements in accuracy and customer experience

The company maintains that AI-based repair recommendations can reduce errors and shorten resolution times, which translates into a better experience for the end customer: fewer delays, fewer back-and-forths, and greater predictability in payment delivery.

From the market’s perspective, Gareth Lodge, principal analyst for global payments at Celent, underscored that a bank’s ability to maximize STP is critical, and that complex inquiries and exception processing consume operational capacity, raise per-transaction costs, and worsen the customer experience. In that context, AI tools aimed at productivity can help tackle the problem.

Operational benefits of OperatorAssist for banks

Reduction in manual investigation time

Among the measurable impacts that Finastra attributes to OperatorAssist is a 20–30% reduction in the time devoted to manual investigation in exception handling.

The logic is that, by automating analysis and suggesting repairs, the operator spends less time on diagnostic tasks and more on closing cases, with less reliance on manual follow-ups and ad hoc reports.

Saving daily hours for users

Finastra claims the tool can save more than 1.5 hours per user per day by speeding up investigations, standardizing resolution steps, and acting as a kind of “virtual expert.” That role also aims to speed up onboarding for new team members, reducing the time until they reach productivity.

Operational KPI Expected day-to-day impact How it is typically measured (practical example) What can bias the measurement
Average exception resolution time Fewer payments “in queue” and less rework Median/percentiles (P50/P90) of time from opening to closing by exception type Mixing simple and complex cases in the same average; priority changes due to volume spikes
% STP (straight-through processing) More payments completed without intervention % of payments that do not generate an operations case / do not require repair Changes in rules, source data quality, ISO 20022 migrations
Reopening / rework rate Fewer back-and-forths with internal areas/customer % of closed cases that are reopened within X days Incomplete documentation, lack of traceability, changes in criteria
Productivity per operator More cases closed per shift without lowering quality Cases closed per hour + quality control (sampling) Incentives that push to “close fast” without addressing the root cause
Time from onboarding to productivity New operators contribute sooner Days/weeks to reach a threshold of cases resolved with quality Differences in prior training, complexity of the payments portfolio

Integration of OperatorAssist with cloud payments platforms

OperatorAssist is offered as an additional capability for customers of Global PAYplus and Payments To Go, expanding Finastra’s cloud payments capabilities with an optional AI layer. The approach, according to the company, is to add assistance within the existing payments hub (not replace it), to speed up the day-to-day work of operations teams when incidents and exceptions arise. The company highlights that the approach is cloud-native and ISO 20022-native, aligned with the modernization of payment infrastructures and the growing standardization of messaging.

In operational terms, this integration aims to make AI available “where the work happens”: within the payments management flow, not as a separate tool.

Operational Integration Assessment
Quick checklist to assess an “ready-to-operate” integration
– Data and traceability: does the hub expose history of the

payment, statuses, and enough events to reconstruct the case without “hunting” for information across multiple systems?
– ISO 20022: are the relevant messages and fields standardized and available for analysis (including variants by scheme/correspondent banking)?
– Roles and permissions: who can accept a suggested repair, who reviews it, and who audits it?
– Workflow: how are queues, priorities, and SLAs routed (by exception type and customer/channel criticality)?
– Operational log: is there a record of the recommendation, the action taken, and the reason (for audit and learning)?
– Deployment and adoption: is there a pilot plan with metrics (time per case, rework, STP) and a checkpoint to adjust rules/processes?

Importance of artificial intelligence in banking operations

The adoption of AI in banking operations is moving from experimental to becoming an efficiency lever. In payments, the value is usually concentrated on three fronts: reduction of manual work, better consistency in decisions, and shorter cycle time in incidents that directly impact costs and customer satisfaction.

OperatorAssist fits into that trend: applying AI not so much to “reinvent” the payment, but to make the operation that supports it more efficient, especially when STP is not possible and the case falls into exception.

Future outlook and adoption of AI in financial services

The launch suggests a clear market direction: more AI tools focused on operational productivity, integrated into core platforms and payment hubs, with impact metrics (time, cost, errors) as the adoption criterion.

As volumes grow, payment schemes accelerate, and the use of ISO 20022 expands, the pressure to resolve exceptions quickly will increase. In that scenario, solutions like OperatorAssist compete to become a de facto standard for operations teams that need to scale without proportionally multiplying headcount.

Signals for applying AI in payments
Signals to watch to decide “when” and “where” to apply AI in payments operations
1) STP gap: if STP stalls, segment by exception type and tackle the highest-volume ones first.
2) Cycle time (P90): prioritize where “long” cases consume capacity and create more customer friction.
3) ISO 20022 maturity: the more consistent the messaging/data, the more predictable automation tends to be.
4) Rework cost: if there are frequent re-openings, focus on traceability and closure quality, notonly speed.
5) Capacity and talent: if the team depends on a few “experts,” an assistant can standardize criteria and speed up onboarding.
6) Measurable operational ROI: define 2–3 KPIs before the pilot (e.g., time per case, rework, STP) and review in short cycles.

Transformation in Payment Handling with Finastra AI OperatorAssist

Innovation in Operational Efficiency

OperatorAssist aims to turn exception management into a more guided and automated process, with concrete promises: more than 1.5 hours saved per user per day, according to Finastra.

Improvements in the Customer Experience

By reducing errors and resolution times, the tool seeks to impact what the customer perceives: more on-time payments, less uncertainty, and faster responses to incidents.

Impact on the Financial Industry

The proposal reinforces a trend: AI applied to operations not only as analytics, but as an integrated assistant in the workflow. If results hold at scale, exception handling—historically expensive and slow—could become a key area of operational differentiation among banks.

Operational Limits and Balances
Limits and trade-offs worth keeping in mind
– Data quality: if the payment history or ISO 20022 fields arrive incomplete, the recommendation may be less reliable and the operator will continue “investigating.”
– Infrequent cases (edge cases): rare or new exceptions (due to scheme/rule changes) often require more human review.
– Human control vs speed: speeding up closures without good logging can increase re-openings; the balance lies in guiding and documenting, not just “solving quickly.”
– Process change: the real benefit depends on adoption (consistent use of the guided flow), queue/SLA adjustments, and coordination between teams.

Transform Your Customer Experience with Suricata Cx

The Definitive Solution for ISPs and Telecommunications Operators

Suricata Cx positions itself as a platform aimed at improving the customer experience in telecommunications environments, with a focus on efficiency and service quality.

Leverage Artificial Intelligence to Optimize Your Operations

AI applied to operations and customer service enables

automate repetitive tasks, speed up response times, and improve service consistency, especially in organizations with a high volume of interactions.

Finastra launches OperatorAssist for AI-powered payments handling, and that approach of AI embedded in the workflow to reduce exceptions, time, and errors reflects exactly the kind of operational automation that at Suricata Cx we apply in telecom: assistants that speed up resolution and keep people in control when the case gets complicated.

From an omnichannel operations and applied automation perspective, the value is usually less in “replacing” teams and more in standardizing diagnostics, reducing rework, and shortening time to resolution, while maintaining traceability and human control in non-routine cases.

Faster and more consistent resolution
Operational example (before/after) to ground the parallel with CX
– Before: an agent receives a “non-routine” case, looks for information across multiple screens, asks another team for context, and the customer waits without clear visibility.
– After (with an assistant in the flow): the case arrives with context, next steps are suggested, and closure is standardized with traceability; the agent spends more time resolving and less time “reconstructing” the problem.
In payments or in telecom, the pattern is the same: when volume grows, sustained improvement usually comes from reducing rework, shortening diagnosis time, and making the resolution process consistent.

This article is based on publicly available information at the time of publication. The cited metrics are results or early estimates and may vary depending on the type of exceptions, volume, and level of operational adoption. In AI products, real-world performance depends heavily on data quality, integration, and process changes, so updates may arise as new information becomes available.