About the public research edition
What FS Atlas covers.
How to read the evidence.
FS Atlas maps AI use cases across financial services and selected adjacent industries. It helps financial-services leaders, practitioners and AI builders discover workflows, compare reported results and identify questions for their own evaluation.
Coverage and counting
The public collection includes banking, wealth and asset management, insurance, payments, markets and financial infrastructure. Selected examples from healthcare, software, legal services, industry and the public sector are included where a workflow could be useful in financial services. That relevance is an inference, not evidence of a financial-services deployment.
| Collection measure | What it means |
|---|---|
| 1,059 catalog records | Preserved use-case entries, including historical records and their later assessments. A record is not necessarily a current production deployment. |
| 1,002 research dossiers | Catalog records with an attached research account: sources, reported stage, uncertainties and review date. A dossier does not establish measured ROI. |
| 17 outcome briefs | Separate accounts of selected reported measures, estimates, studies, models or testimonials, with context and limitations. Some describe the same initiatives as catalog records. |
| 1,035 discovery initiatives | The combined overview after explicit catalog-to-brief overlaps are consolidated and legacy records assessed as unsupported or duplicate are excluded. |
The categories above overlap and must not be added together. The library preserves original records for transparency; the discovery view offers a more selective exploration set. 107 legacy records currently have an explicit assessment. This is bounded public-source research, not an exhaustive census or a ranking of institutions.
Sources and limitations stay visible
Research uses issuer disclosures, annual reports, filings, product announcements, named customer case studies, public studies and secondary reporting. Source registers identify the publisher, date and access level where known. Full-text review, indexed text, secondary-only access and blocked sources are distinguished; a citation alone is not proof of a claim.
- Company-reported outcomes retain their original scope and period. A number in a company filing or report is not automatically an independently audited AI result.
- Vendor case studies are supplier-hosted accounts. Customer naming and reported production use do not establish independent validation.
- Estimates, forecasts, models and testimonials carry distinct labels. They should not be presented as measured, realized benefits.
- Deployment stage describes what the source supports at its own date. An announcement, pilot and scaled deployment are different states.
- Planning assessments of effort, cost and dependencies are labeled inferences for a bounded pilot. They are not disclosed project budgets, implementation quotes or delivery commitments.
Reported cost, time, satisfaction and risk measures often have different baselines, denominators and attribution methods. Atlas keeps those limitations with the result rather than combining them into a single return-on-investment score. The research briefing shows selected evidence in context.
What the research dates mean
Latest research round: . The published snapshot was generated on November 8, 2026 at 11:18 PM ET.
Individual dossier review dates span Oct 26, 2026 to Nov 8, 2026. The latest review date among the 17 separate outcome briefs is Oct 26, 2026. A new research round does not mean every existing record or brief has been rechecked.
A review date records when that research account was prepared or updated. A source publication date records when the underlying report appeared; a deployment or measurement date may be different again. Historical sources do not prove that the same system remains in use, at the same scope, today.
The public site is a read-only editorial snapshot. It is refreshed through new research releases and is not a continuously operating monitoring service. Refer to each case for its own review date, source dates and unresolved questions.
What remains uncertain
Public evidence favors organizations and vendors that publish. Smaller firms, private deployments and less accessible languages or regions can be underrepresented. Missing evidence is not evidence of non-use. Source access and coverage gaps are recorded in the collection notes.
Public Reddit and X discussions inform selected practitioner reality checks. These are purposive samples: identities, employment and underlying events are generally unverified. Comments can suggest failure modes to test; they do not establish prevalence, causal effects or representative customer sentiment.
Cross-industry results are presented with transfer limits. Different regulation, data permissions, review obligations, error tolerance and integration costs may change what is practical in a financial institution.
Publisher and editorial approach
FS Atlas is published by Aluda Saliashvili. Research, synthesis and categorization are AI-assisted. The site does not represent independent audit, institution endorsement or expert certification of the listed deployments.
Source links, corrections, legacy assessments and recorded uncertainties are preserved so readers can inspect the basis of an account. For a decision, return to the original source and validate the workflow, controls and economics in your own operating context.
Open the opportunity tree to explore themes and workflows, or browse the full library to inspect individual records.