The news
Finastra, the banking software company backed by Vista Equity Partners, launched Repair Recommendations on September 29, 2026 at the Sibos banking conference in Miami, adding AI-suggested fixes for failed or flagged payments to its AI OperatorAssist tool for bank operations teams.
When a payment breaks a rule, carries bad data or cannot be matched, it drops out of automated processing and typically lands in an exception queue, where specialists investigate and repair it manually. Finastra said the new feature will identify payment discrepancies, analyse their underlying causes and recommend corrections validated against the rules of payment schemes and networks, naming Swift, Fedwire, SEPA, UPI and Nexus.
According to Finastra, the recommendations are designed to support human decisions, with each action subject to review and approval. The company said the feature aims to cut the time wire-room operators, exception and repair specialists and technical support staff spend resolving exceptions, reduce error rates and help new staff get up to speed with less reliance on scarce experts.
Repair Recommendations works with Finastra's payment hubs, Global PAYplus and Payments To Go, and extends OperatorAssist, which the company launched on March 5, 2026. At that launch, Finastra said OperatorAssist reduced manual investigation time by 20-30%, saving users more than 1.5 hours a day, and that early results pointed to efficiency gains upwards of 20%. The September release gave no new performance figures, pricing or customer names.
Finastra framed the problem in broad terms, saying payment processing errors and disruptions cost the industry billions a year; it did not give a source for the estimate. Barry Rodrigues, Finastra's executive vice president for payments, said AI can help deliver immediate value as payment volumes increase. Robin LoGiudice, a strategic adviser at research firm Datos Insights, said in the release that manual processes remain a major operational risk for financial institutions.
Trade publication PYMNTS reported the launch the same day, noting that all actions require human review before they are carried out.
The numbers
- Payment schemes named for rule validation
- 5 (Swift, Fedwire, SEPA, UPI, Nexus)
- Manual investigation time cut by OperatorAssist (Finastra, March 2026)
- 20-30%
- Time saved per user per day (Finastra, March 2026)
- More than 1.5 hours
- Efficiency gains from early results (Finastra, March 2026)
- Upwards of 20%
Why CEOs should care
For bank COOs and heads of payments, exception handling is a cost that rarely shows up in a headline but grows with every new payment rail. Instant and cross-border payments leave less time to fix a problem before a customer notices. When evaluating Finastra's feature or rivals, ask for measured results from live banks, not just the March estimates: how many exceptions per day it handled, what share of its suggestions staff accepted unchanged, and how often an accepted fix later failed.
Chief risk and compliance officers should look closely at the approval model. Finastra says every action is reviewed by a person, which keeps accountability with the bank but also means the savings depend on how quickly staff can check a suggestion. Ask how the tool logs the reason for each recommendation, whether examiners can audit it, and whether rule updates for Swift, Fedwire, SEPA and other schemes are pushed automatically when those networks change their rules.
CFOs and boards should treat this as a staffing and resilience question. Finastra pitches the feature as a way to onboard new staff faster and depend less on a few experts. That is valuable, but banks should also keep enough trained people to repair payments manually if the AI tool is unavailable or wrong during a volume spike.
The bigger picture
Finastra's release ties the feature directly to rising demand for real-time cross-border payments, infrastructure modernisation and digital commerce. When OperatorAssist launched in March, Gareth Lodge, principal analyst at research firm Celent, said in Finastra's release that complex inquiries and exception processing can take a significant amount of capacity from bank operations. Both launches target back-office work that grows as banks connect to more payment schemes.
The approach also reflects a cautious middle path. Rather than letting AI repair payments on its own, Finastra keeps a human approval step for every action, a design banks and regulators are likely to prefer while the technology is new in money movement.
What’s next
Watch for the first banks to report results from Repair Recommendations in production, and whether Finastra later moves some low-risk fixes to automatic approval. Competing payment hub vendors are likely to answer with their own AI exception tools as instant payment volumes grow.
What “Fact-checked” means
Fact-checking means testing a story’s facts against the evidence before it is published. This story went through at least two separate checks before this version was published.
- What we checked
- Its names, figures, dates, job titles, quotes and who said what were checked against the story’s sources, including its main source where it could be opened. The headline was checked for accuracy and overstatement.
- How
- A first check reviewed the whole story. If it passed, a second, skeptical check went back to the sources to look for mistakes in the most important facts. If a check flagged the story, it was edited to fix the problems found, and a separate re-check then reviewed the whole story again.
- Who
- The checks are made by our newsroom, as steps kept separate from the writing, under rules set by our editor, Hussein Mukhtar. A story the checks still flag is held for the editor, who decides whether it is fixed, published or dropped.
- If something is wrong
- “Fact-checked” does not mean error-free. If a material error is found after publication, we correct the story and add a note saying what changed. Report an error
