Finastra

Strip every model and Loan IQ still services syndicated loans, LaserPro still generates documents as it has since 1986 and Mortgagebot still originates mortgages. Seven separately branded AI products are real, yet each is an assistant, search tool or ingestion aid positioned beside the systems of record.

Humans in the loop is the stated posture, and every named capability assists, searches, trains, translates or ingests without executing a transaction. What verifies an extracted credit agreement term before it drives interest calculations on a live facility is never described.

A white paper on generative AI and seven branded capabilities sit beside no accuracy or error rate for document classification, term extraction or translation. No validation method, drift monitoring or external assessment describes how any output is checked.

ING, Maybank and United Bank appear as named case studies or interviews, and the platform services roughly seventy percent of global syndicated loan volume. One customer says nearly ninety five percent of its portfolio sits on the platform, but the quote is unattributed and no named case carries a number.

Loan documents and portfolio positions of competing lenders share one platform, and nothing states whether customer data trains any model or is pooled across institutions. With a third party model provider inside the extraction flow, what that supplier is bound to matters as much as the vendor's own terms, and neither is published.

Credit agreements, among a corporate borrower's commercially sensitive documents, now pass through a partner supplied extraction service at onboarding. Only a general website privacy policy exists, with no retention, deletion or subprocessor terms covering what either party keeps of that content.

The one published certification is a 2018 announcement covering one Swiss service bureau, entity scoped and eight years old. No current group wide credential is published on product pages or in the footer, for a platform holding loan positions across much of syndicated lending.

Servicing a large share of global syndicated lending makes the platform operationally close to market infrastructure, but the company claims no license or supervised standing. The documentation product's regulatory content and fifty state attorney network are a compliance service sold to buyers, not standing held by the firm.

A named chief AI officer sits beside no published AI governance framework, fairness testing or position on any AI regulation. The assistive capabilities score no borrower, which limits fairness exposure, but mortgage origination and documentation reach consumer lending directly with nothing addressing that path.

The risk is a wrongly captured instrument rather than a declined borrower: an extracted margin, maturity or covenant applied across a syndicate. Nothing divides responsibility among vendor, extraction partner and agent bank, and the documentation warranty covers regulatory content, not model output.

Marketnode is named as the supplier of large language model document automation behind loan onboarding, with the integration point identified by module. Naming the partner is not naming the model, and no base model, provider or version appears for the partner capability or the company's own products.

Loan IQ is the system of record for the loan itself, on a single data model, and Loan IQ Nexus is sold specifically for connectivity across internal and external market platforms. A universal banking core, a payments hub and a partner marketplace with third party lending apps surround it.

The lending platform runs on premises or in container based cloud, with a named managed hosting service and a cloud edition of the documentation product. No cloud provider, region or transfer mechanism is named for loan and borrower data spanning Europe, the United States and Asia.

A syndicated lending platform, a documentation product for community institutions and a managed hosting service are three very different commercial shapes, and none has a published price, tier or licensing model. Every route ends at a contact form.

Nine institution segments are named explicitly, from community banks and credit unions to corporates and fintechs, rather than inferred from logos. Reach runs from more than three thousand institutions on the documentation product to 21 of the top 25 syndicated lenders, with private credit lenders and asset servicers beyond banks.