Platform
How a name becomes a defensible report
PRSM is a managed investigation workflow delivered as a product. Source checks feed human analysts; evidence backs every finding; QA gates delivery; and the work compounds into reusable intelligence.
The workflow
One flow, from intake to reusable intelligence
The public service narrative — identity, OSINT, AML, media, delivery — runs on this operational backbone. This is the source of truth for what actually happens to a case.
- 01
Intake
A client submits one or more people or entities. Submitted values are preserved exactly, and a case is created — nothing is silently normalized away.
- 02
Triage & assignment
Admins triage the request, set priority and due date, and assign an investigator. Client-safe status labels hide internal operational detail.
- 03
Investigator review
Investigators work the subject through a checklist: source checks, evidence capture, candidate findings, and relationship mapping.
- 04
Evidence before findings
A finding cannot move toward QA until it cites evidence — source, timestamp, snapshot, and matching logic are preserved, not summarized away.
- 05
Relationship graph
People and entities are linked with typed, confidence-rated edges so indirect and transitive exposure can be reasoned about — and shown.
- 06
QA & approval
A QA reviewer approves, holds, or requests revisions on each section. Approval is a gate, not a formality.
- 07
Secure delivery
The approved report is published to a tenant-scoped client portal. Internal notes are never selected into the client-facing report.
- 08
Reusable knowledge
Confirmed records, aliases, identifiers, evidence, findings, and relationships remain available — so the next case starts ahead of where the last one began.
Report anatomy
Structured to be read by people who carry the decision
A PRSM report synthesizes public, governmental, and open-source material into decision support — built to hold up in front of a board, a regulator, or opposing counsel.
- 01
Executive summary
Decision-support overview, not a raw dump.
- 02
Subject & identity
Profile, aliases, identifiers, transliterations.
- 03
Risk category table
Each category explained in plain English.
- 04
Direct findings
What the subject is directly tied to, with evidence.
- 05
Indirect exposure
Relationship-path risk, with hop counts.
- 06
Relationship map
How the exposure actually connects.
- 07
Evidence summary
Every claim traceable to a source record.
- 08
Source coverage
What was checked — and the limits of each source.
- 09
Unresolved questions
What is uncertain, stated plainly.
- 10
Recommended follow-up
The next screening worth running.
Direct vs. indirect
The distinction commodity screening blurs
Most counterparty risk lives in relationships, not in a single list hit. PRSM keeps direct findings, indirect exposure, and transitive paths separate — and labels each one.
Direct
The subject itself appears in a source, list, court record, registry, or enforcement context.
Indirect
The subject is connected through an ownership, trade, directorship, or associated-entity relationship.
Transitive
A risk path that spans more than one hop — surfaced with its hop count, never flattened into a direct match.
Risk categories, explained
Short codes like SAN or JUR never travel without a plain-English explanation.
- SAN Sanctions Direct listing or sanctions-network adjacency.
- RES Restricted entity Export-control, watchlist, or restricted-party signals.
- JUR Jurisdiction Exposure to higher-risk jurisdictions — without implying misconduct.
- REP Reputational Adverse media and public-record reputational signals.
- GOV Governance Ownership opacity, control, and governance concerns.
- LEG Legal & court Litigation, court actions, and legal proceedings.
- ENF Enforcement Regulatory or enforcement activity.
- IND Indirect exposure Transitive risk through ownership, trade, or association.
Calibrated confidence
Findings carry deliberate language. “No adverse finding found” is not “no risk,” and a possible match is never written up as a confirmed one.
- Confirmed match Evidence directly supports the match.
- Probable match Strong but not conclusive support.
- Possible match A candidate signal requiring more work.
- Unresolved signal Surfaced, not yet adjudicated.
- Rejected match Reviewed and ruled out.
- No adverse finding found Not the same as “no risk.”
- No sufficient public signal The public record is thin here.
Reusable intelligence
The investigation becomes an asset
PRSM stores investigation output as structured records — not just a final PDF. Confirmed entities and relationships make every subsequent case faster and more consistent.
Canonical person / entity
Reusable internal records, linked only after human confirmation.
Aliases & identifiers
Alternate names, transliterations, registry IDs, DOBs, company numbers.
Evidence item
Source-backed document, screenshot, URL, registry record, or upload.
Finding
An investigator-authored statement, tied to its evidence.
Relationship
A typed edge between parties, with confidence and source evidence.
Report
The customer-facing product, assembled from approved findings.
Canonical linking and record merging are never automatic. Human confirmation is part of PRSM's trust posture.
Automation boundary
AI assists the analyst. It never replaces the judgment.
Automation earns its place by saving investigators time — not by quietly making the call.
- Summarize evidence for an investigator to review.
- Extract names, dates, jurisdictions, and relationship candidates.
- Suggest search ideas or draft report language.
Allowed
- Approve findings or decide a match.
- Deliver a report on its own.
- Turn a weak signal into a confirmed claim.
Never
See it on a real case
Walk through the platform with us
We’ll show the workflow end to end on a representative case and scope what a review of your subject would involve.
Human-reviewed. Evidence-backed. Handled with tenant-scoped confidentiality.