Skip to main content
Forensic Analysis

How to Detect Tender Irregularities with Forensic Analysis on Tenders-SA

A practical guide to using the Tenders-SA Forensic Analysis workbench for supplier audits, department scans, pair analysis, and AI-generated case narratives to identify procurement irregularities.

How to Detect Tender Irregularities with Forensic Analysis on Tenders-SA

South African public procurement spends hundreds of billions of rand annually across national, provincial, and municipal government. With this volume of expenditure comes the risk of procurement irregularities: awards to non-compliant suppliers, unusual clustering below threshold levels, director conflicts across multiple awarded companies, and contracts going to deregistered or restricted entities.

The Tenders-SA Forensic Analysis tool at /tools/forensic-analysis is an investigative workbench designed to detect these patterns. It combines public tender award data, CIPC company registry data, the National Treasury OCPO restricted supplier list, and competitive intelligence metrics into a single research interface. This guide walks through the available tools and workflows.

What the Forensic Analysis Tool Does

The forensic analysis engine applies algorithmic detection to identify patterns in procurement data that may warrant further investigation. It is structured around ten detection flags, each targeting a specific type of irregularity or risk pattern. The tool scores suppliers and departments on a 0 to 100 risk scale and produces named, traceable findings with supporting data.

The tool is explicitly positioned as a pattern detection and triage system, not a legal judgment engine. Its findings identify statistical unusualness and named data patterns for review. They do not establish wrongdoing and require source verification.

Starting with the Forensic Workbench

The workbench landing page offers three primary entry points: Supplier Audit for investigating a specific company, Department Scan for reviewing a buying organisation's award patterns, and the search workbench for browsing and filtering entities.

The search workbench at /tools/forensic-analysis/search supports two modes: Supplier and Buyer. It provides filters for province, buyer type, municipality, category, award year range, value band, award count band, enterprise type, B-BBEE level, OCPO status, and intelligence exposure. Results are sortable by highest value, most awards, most recent, highest risk, and highest concentration.

Free tier users see preview results with up to eight rows of data and capped flag information. Full access to all rows, advanced filters, and detailed reports requires an active Pro subscription.

Supplier Forensic Report

Clicking through to a supplier forensic report at /tools/forensic-analysis/supplier/{slug} provides a consolidated view of the supplier's risk profile. The report includes:

  • CIPC enrichment data: registration number, status, company type, compliance score, physical address, and director count.
  • Computed forensic risk score from 0 to 100 with named flag contributors explaining what drove the score.
  • Award timeline showing up to 50 awards for Pro subscribers or 10 for free users.
  • Cross-links to the Company Intelligence profile at /tools/company-intelligence/{slug} for award history and enrichment context.

The forensic risk score is displayed with a severity band: Clear (0-24), Watch (25-49), Elevated (50-74), or High Concern (75-100). Below the score, a breakdown section lists each contributing flag with its severity, description, and supporting data.

Department Forensic Report

The department-level report at /tools/forensic-analysis/department/{slug} provides a buyer-wide perspective. It shows:

  • Award statistics: total awards count and total known value.
  • Top 20 suppliers ranked by value and award count.
  • Herfindahl-Hirschman Index (HHI) measuring market concentration in the department's awards.
  • Composite department risk score aggregating all supplier-level flags.
  • Detection results for director overlaps, address clusters, new entrant large awards, and threshold clusters.

The department report is particularly useful for auditors and oversight officials who need to assess the overall health of a procuring entity's award practices.

Pair Scan: Analysing Buyer-Supplier Relationships

The Pair Scan feature addresses a specific investigative question: is there anything unusual about the relationship between this supplier and this buyer? It returns:

  • Total awards and total value between the buyer-supplier pair.
  • Average award value, latest award date, and categories awarded.
  • Threshold signals: awards near R500,000, R1 million, and R2 million procurement threshold levels.
  • Supplier OCPO restricted status, if applicable.
  • Procurement intelligence context, matching relevant news or intelligence items.

The pair scan is designed for investigators who need to quickly assess whether a specific buyer-supplier relationship warrants deeper investigation.

Entity Comparison

The Compare Tray allows you to select two to five entities from the workbench and compare them side by side. Comparison dimensions include total awards and value, average award value, latest award date, provinces of operation, top counterpart organisations, B-BBEE level, OCPO status, and competitive concentration metrics all displayed on a single page.

Restricted Supplier Verification

The Restricted Suppliers browser at /tools/restricted-suppliers provides direct access to the National Treasury OCPO restricted supplier list. The list is synced daily from the government PDF and includes supplier names, registration numbers, restriction types and reasons, restriction periods, authorising bodies, and active status. The system also cross-references restricted suppliers against the World Bank and AfDB debarment lists.

The date-aware restriction overlay is one of the most powerful features. It layers award timing over OCPO records to distinguish between suppliers that were restricted during the period of an award (the strongest signal), suppliers that are currently restricted but had no award overlap, and suppliers with expired restriction history.

Restricted supplier status is displayed through a series of visual components: an inline badge on search results, a full-width banner on company profiles, and hover tooltips with restriction details. A clear green badge indicates no restriction found; an amber badge indicates an expired restriction; and a red badge indicates an active restriction, possibly overlapping with awards.

AI Narrative Generation

For Pro subscribers, the AI narrative generator creates a structured audit summary from the forensic findings. The AI is prompted to use AGSA-style language with cautious formulations such as 'data indicates,' 'pattern suggests,' and 'warrants further investigation.' The generated narrative is three to four paragraphs covering key findings and recommended actions.

The AI narrative is intended as a starting point for investigation, not as a final determination. The system prompt explicitly prohibits terms such as corruption, bribery, and criminal, substituting irregular, anomalous, and other investigatory language.

Practical Investigation Workflow

A typical forensic investigation using the platform follows these stages:

  1. Triage: Use the workbench search to identify suppliers or departments that warrant attention, filtered by risk band or OCPO status.
  2. Deep dive: Open supplier forensic reports to review the risk score breakdown, flag details, and award timeline.
  3. Cross-reference: Check the restricted supplier browser for any OCPO records and use the date-aware overlay to assess whether restrictions overlap with awards.
  4. Contextualise: Use the intelligence feed at /intelligence to check if the department or supplier has been mentioned in recent procurement intelligence items.
  5. Document: Generate an AI narrative as a starting brief, and use the evidence timeline to compile a chronological view of events.
  6. Escalate: Export the findings for further investigation or formal procurement review.

Limitations and Disclaimers

The forensic analysis tool has several important limitations. The CIPC enrichment that feeds many of the forensic flags covers only a portion of the awarded supplier universe; the system is transitioning toward procurement-native scoring that relies primarily on award data. The department-level buyer report has a known technical issue with slug resolution that can produce empty results in some cases, and this has been identified for a planned fix.

All forensic findings are accompanied by a standard disclaimer: 'These findings identify statistical unusualness and named data patterns for review. They do not establish wrongdoing and require source verification.' The tool is an investigative aid, not a substitute for qualified forensic audit or legal review.

Summary

The Forensic Analysis workbench provides a structured investigative environment for detecting procurement irregularities. Its combination of pattern detection, restricted supplier verification, buyer-supplier pair analysis, and entity comparison tools makes it suitable for auditors, procurement officials, investigators, and suppliers conducting due diligence. Access the workbench at /tools/forensic-analysis and the restricted supplier browser at /tools/restricted-suppliers to begin an investigation.

Tags

Forensic AnalysisTender IrregularitiesProcurement Fraud DetectionSupplier AuditOCPO Restricted SuppliersTender Investigation
AI-Powered Matching
Never Miss a Perfect Tender Again
Our AI analyzes thousands of tenders and finds the ones YOUR company can actually win
AI Match Scoring for every tender
Instant alerts for 85%+ matches
B-BBEE level optimization
Document readiness checks

Share this article

How to Detect Tender Irregularities with Forensic Analysis on Tenders-SA

A practical guide to using the Tenders-SA Forensic Analysis workbench for supplier audits, department scans, pair analysis, and AI-generated case narratives to identify procurement irregularities.

https://www.tenders-sa.org/blog/detect-tender-irregularities-forensic-analysis