{"id":947,"date":"2026-11-03T12:12:52","date_gmt":"2026-11-03T01:12:52","guid":{"rendered":"https:\/\/nexiant.ai\/resources\/blogs\/?p=947"},"modified":"2026-11-11T12:17:20","modified_gmt":"2026-11-11T01:17:20","slug":"when-rules-and-ai-scores-disagree","status":"publish","type":"post","link":"https:\/\/nexiant.ai\/resources\/blogs\/when-rules-and-ai-scores-disagree\/","title":{"rendered":"What Should Happen When a Rule and an AI Risk Score Disagree?"},"content":{"rendered":"\n<style>\n  .nx-blog * { box-sizing: border-box; margin: 0; padding: 0; }\n  .nx-blog { font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, sans-serif; font-size: 16px; line-height: 1.75; color: #1a1a2e; max-width: 820px; margin: 0 auto; }\n  .nx-blog h2 { font-size: 1.55rem; font-weight: 700; color: #00184C; margin: 2.5rem 0 0.75rem; padding-bottom: 0.4rem; border-bottom: 3px solid #073EA1; }\n  .nx-blog h3 { font-size: 1.15rem; font-weight: 700; color: #073EA1; margin: 1.75rem 0 0.5rem; }\n  .nx-blog h4 { font-size: 0.98rem; font-weight: 700; color: #00184C; margin: 1.25rem 0 0.35rem; }\n  .nx-blog p { margin-bottom: 1rem; }\n  .nx-blog ul, .nx-blog ol { margin: 0.5rem 0 1rem 1.4rem; }\n  .nx-blog li { margin-bottom: 0.4rem; }\n  .nx-blog strong { color: #00184C; }\n\n  .nx-hero { background: linear-gradient(135deg, #00184C 0%, #073EA1 100%); color: #fff; border-radius: 12px; padding: 1.75rem 2rem; margin-bottom: 2rem; }\n  .nx-hero .nx-tag { display: inline-block; background: rgba(255,255,255,0.15); color: #AEC9FF; font-size: 0.75rem; font-weight: 600; text-transform: uppercase; letter-spacing: 0.08em; padding: 4px 12px; border-radius: 20px; margin-bottom: 0.6rem; }\n  .nx-hero .nx-meta { font-size: 0.95rem; color: #AEC9FF; margin: 0; }\n\n  .nx-callout { border-left: 4px solid #073EA1; background: #f0f4ff; border-radius: 0 8px 8px 0; padding: 1rem 1.25rem; margin: 1.5rem 0; }\n  .nx-callout.nx-callout--warning { border-left-color: #A30000; background: #fff5f5; }\n  .nx-callout .nx-callout-title { font-size: 0.8rem; font-weight: 700; text-transform: uppercase; letter-spacing: 0.07em; color: #073EA1; margin-bottom: 0.4rem; }\n  .nx-callout.nx-callout--warning .nx-callout-title { color: #A30000; }\n  .nx-callout p { margin: 0; font-size: 0.95rem; color: #1a1a2e; }\n\n  .nx-media { margin: 1.5rem 0 1.75rem; }\n  .nx-media img { width: 100%; height: auto; display: block; border-radius: 12px; border: 1px solid #d0daf5; background: #f5f8ff; }\n  .nx-media .nx-caption { font-size: 0.82rem; color: #666; text-align: center; margin-top: 0.6rem; line-height: 1.5; font-style: italic; }\n\n  .nx-flow { display: grid; grid-template-columns: repeat(4, 1fr); gap: 12px; margin: 1.5rem 0 1.75rem; position: relative; }\n  .nx-flow-step { background: #fff; border: 1px solid #d0daf5; border-radius: 12px; padding: 1rem; text-align: center; position: relative; }\n  .nx-flow-step::after { content: \"\u2192\"; position: absolute; right: -14px; top: 50%; transform: translateY(-50%); color: #073EA1; font-weight: 700; font-size: 1.2rem; }\n  .nx-flow-step:last-child::after { display: none; }\n  .nx-flow-number { width: 30px; height: 30px; border-radius: 50%; background: #073EA1; color: #fff; font-size: 0.8rem; font-weight: 700; display: flex; align-items: center; justify-content: center; margin: 0 auto 0.6rem; }\n  .nx-flow-step h4 { font-size: 0.9rem; color: #00184C; margin-bottom: 0.35rem; }\n  .nx-flow-step p { font-size: 0.8rem; color: #555; margin: 0; line-height: 1.5; }\n\n  .nx-grid { display: grid; grid-template-columns: repeat(auto-fit, minmax(185px, 1fr)); gap: 12px; margin: 1.25rem 0 1.75rem; }\n  .nx-card { background: #fff; border: 1px solid #d0daf5; border-radius: 10px; padding: 1rem; }\n  .nx-card-icon { width: 36px; height: 36px; border-radius: 8px; background: #073EA1; display: flex; align-items: center; justify-content: center; margin-bottom: 0.6rem; }\n  .nx-card-icon svg { width: 18px; height: 18px; fill: #fff; }\n  .nx-card h4 { font-size: 0.88rem; font-weight: 700; color: #00184C; margin-bottom: 0.2rem; }\n  .nx-card p { font-size: 0.8rem; color: #555; margin: 0; line-height: 1.5; }\n\n  .nx-table-wrap { overflow-x: auto; margin: 1.25rem 0 1.75rem; }\n  .nx-table { width: 100%; border-collapse: collapse; font-size: 0.9rem; }\n  .nx-table thead tr { background: #00184C; color: #fff; }\n  .nx-table th { text-align: left; padding: 10px 14px; font-weight: 600; }\n  .nx-table td { padding: 9px 14px; border-bottom: 1px solid #e0e7f5; color: #1a1a2e; vertical-align: top; }\n  .nx-table tbody tr:nth-child(even) { background: #f5f8ff; }\n  .nx-badge { display: inline-block; font-size: 0.73rem; font-weight: 600; padding: 2px 9px; border-radius: 20px; }\n  .nx-badge--red { background: #fde8e8; color: #A30000; }\n  .nx-badge--blue { background: #EEF2FF; color: #073EA1; }\n\n  .nx-signal-panel { background: #f5f8ff; border: 1px solid #d0daf5; border-radius: 12px; padding: 1.25rem; margin: 1.5rem 0 1.75rem; }\n  .nx-signal-panel h3 { margin-top: 0; color: #00184C; }\n  .nx-signal-grid { display: grid; grid-template-columns: repeat(auto-fit, minmax(180px, 1fr)); gap: 10px; margin-top: 1rem; }\n  .nx-signal { background: #fff; border-radius: 10px; padding: 0.9rem; border-left: 4px solid #073EA1; }\n  .nx-signal strong { display: block; font-size: 0.88rem; margin-bottom: 0.25rem; }\n  .nx-signal span { font-size: 0.8rem; color: #555; line-height: 1.5; }\n\n  .nx-scenario { background: linear-gradient(135deg, #00184C 0%, #073EA1 100%); color: #fff; border-radius: 12px; padding: 1.5rem; margin: 1.75rem 0; }\n  .nx-scenario h3 { color: #fff; margin-top: 0; margin-bottom: 0.6rem; }\n  .nx-scenario p { color: #E7EFFF; margin-bottom: 0.8rem; }\n  .nx-scenario ul { margin-bottom: 0; }\n  .nx-scenario li { color: #E7EFFF; }\n\n  .nx-stack { margin: 1.5rem 0 1.75rem; }\n  .nx-stack-layer { display: flex; align-items: flex-start; gap: 14px; background: #fff; border: 1px solid #d0daf5; border-radius: 12px; padding: 1rem; margin-bottom: 10px; }\n  .nx-stack-icon { width: 38px; height: 38px; border-radius: 10px; background: #073EA1; color: #fff; font-weight: 700; font-size: 0.85rem; display: flex; align-items: center; justify-content: center; flex-shrink: 0; }\n  .nx-stack-layer h4 { margin: 0 0 0.25rem; color: #00184C; font-size: 0.95rem; }\n  .nx-stack-layer p { margin: 0; font-size: 0.86rem; color: #444; line-height: 1.6; }\n\n  .nx-obligations { margin: 1.25rem 0 1.75rem; }\n  .nx-obl-item { background: #fff; border: 1px solid #d0daf5; border-left: 4px solid #073EA1; border-radius: 0 10px 10px 0; padding: 1rem 1.25rem; margin-bottom: 10px; }\n  .nx-obl-item h4 { font-size: 0.93rem; font-weight: 700; color: #00184C; margin-bottom: 0.3rem; }\n  .nx-obl-item p { font-size: 0.87rem; color: #444; margin: 0; line-height: 1.6; }\n\n  .nx-inline-link { color: #073EA1; text-decoration: underline; font-weight: 600; }\n  .nx-inline-link:hover { color: #00184C; }\n\n  .nx-faq { margin: 1.25rem 0 1.75rem; }\n  .nx-faq-item { border: 1px solid #d0daf5; border-radius: 8px; margin-bottom: 8px; overflow: hidden; }\n  .nx-faq-q { width: 100%; background: #fff; border: none; text-align: left; padding: 1rem 1.25rem; font-size: 0.95rem; font-weight: 600; color: #00184C; cursor: pointer; display: flex; justify-content: space-between; align-items: center; gap: 1rem; }\n  .nx-faq-q:hover { background: #f5f8ff; }\n  .nx-faq-q .nx-chevron { flex-shrink: 0; width: 20px; height: 20px; border-radius: 50%; background: #EEF2FF; display: flex; align-items: center; justify-content: center; transition: transform 0.25s; }\n  .nx-faq-q .nx-chevron svg { width: 10px; height: 10px; stroke: #073EA1; fill: none; }\n  .nx-faq-q[aria-expanded=\"true\"] .nx-chevron { transform: rotate(180deg); background: #073EA1; }\n  .nx-faq-q[aria-expanded=\"true\"] .nx-chevron svg { stroke: #fff; }\n  .nx-faq-a { display: none; padding: 0 1.25rem 1rem; font-size: 0.92rem; color: #333; line-height: 1.7; background: #fff; }\n  .nx-faq-a.open { display: block; }\n\n  .nx-cta { background: linear-gradient(135deg, #00184C 0%, #073EA1 100%); border-radius: 12px; padding: 2rem; text-align: center; margin-top: 2.5rem; }\n  .nx-cta h3 { color: #fff; font-size: 1.3rem; font-weight: 700; margin-bottom: 0.5rem; }\n  .nx-cta p { color: #AEC9FF; font-size: 0.95rem; margin-bottom: 1.25rem; }\n  .nx-cta a { display: inline-block; background: #E11A1A; color: #fff; font-weight: 700; font-size: 0.95rem; padding: 0.7rem 1.8rem; border-radius: 6px; text-decoration: none; transition: background 0.2s; }\n  .nx-cta a:hover { background: #A30000; }\n  .nx-divider { border: none; border-top: 1px solid #e0e7f5; margin: 2rem 0; }\n  .nx-disclaimer { font-size: 0.8rem; color: #888; font-style: italic; text-align: center; margin-top: 1.5rem; }\n\n  @media (max-width: 720px) {\n    .nx-flow { grid-template-columns: 1fr; }\n    .nx-flow-step::after { content: \"\u2193\"; right: 50%; top: auto; bottom: -18px; transform: translateX(50%); }\n    .nx-stack-layer { flex-direction: column; }\n  }\n<\/style>\n\n<div class=\"nx-blog\">\n\n  <!-- HERO -->\n  <div class=\"nx-hero\">\n    <span class=\"nx-tag\">FraudShield AI Engine &nbsp;\u00b7&nbsp; November 2026 &nbsp;\u00b7&nbsp; Australia &amp; International<\/span>\n    <p class=\"nx-meta\">Why different risk signals should be explained and reviewed rather than letting one overwrite the other.<\/p>\n  <\/div>\n\n  <!-- INTRODUCTION -->\n  <p><strong>A rule-based monitoring system<\/strong> can flag a transaction while an accompanying AI score appears relatively low. The reverse is also possible: no rule triggers, yet the AI score is high. These outcomes can look contradictory, but they may actually provide useful information about how two different signals evaluate the same transaction.<\/p>\n  <p>The important question is <strong>how an investigator should interpret disagreement without letting one signal erase the other<\/strong>.<\/p>\n\n  <div class=\"nx-media\">\n    <img decoding=\"async\" src=\"data:image\/svg+xml,%3Csvg xmlns='http:\/\/www.w3.org\/2000\/svg' width='820' height='460' viewBox='0 0 820 460'%3E%3Cdefs%3E%3ClinearGradient id='bg' x1='0' y1='0' x2='1' y2='1'%3E%3Cstop offset='0' stop-color='%2300184C'\/%3E%3Cstop offset='1' stop-color='%23073EA1'\/%3E%3C\/linearGradient%3E%3C\/defs%3E%3Crect width='820' height='460' rx='22' fill='url(%23bg)'\/%3E%3Ctext x='410' y='70' text-anchor='middle' font-family='Arial, sans-serif' font-size='30' font-weight='700' fill='%23ffffff'%3EWhen Rules and AI Scores Disagree%3C\/text%3E%3Crect x='86' y='132' width='165' height='92' rx='18' fill='%23ffffff' opacity='0.92'\/%3E%3Crect x='328' y='132' width='165' height='92' rx='18' fill='%23ffffff' opacity='0.88'\/%3E%3Crect x='570' y='132' width='165' height='92' rx='18' fill='%23ffffff' opacity='0.84'\/%3E%3Crect x='207' y='272' width='165' height='92' rx='18' fill='%23ffffff' opacity='0.86'\/%3E%3Crect x='449' y='272' width='165' height='92' rx='18' fill='%23ffffff' opacity='0.9'\/%3E%3Ccircle cx='168' cy='162' r='16' fill='%23073EA1'\/%3E%3Ccircle cx='410' cy='162' r='16' fill='%23073EA1'\/%3E%3Ccircle cx='652' cy='162' r='16' fill='%23073EA1'\/%3E%3Ccircle cx='290' cy='302' r='16' fill='%23073EA1'\/%3E%3Ccircle cx='531' cy='302' r='16' fill='%23073EA1'\/%3E%3Ctext x='168' y='198' text-anchor='middle' font-family='Arial, sans-serif' font-size='17' font-weight='700' fill='%2300184C'%3ERule result%3C\/text%3E%3Ctext x='410' y='198' text-anchor='middle' font-family='Arial, sans-serif' font-size='17' font-weight='700' fill='%2300184C'%3EAI score%3C\/text%3E%3Ctext x='652' y='198' text-anchor='middle' font-family='Arial, sans-serif' font-size='17' font-weight='700' fill='%2300184C'%3EExplanation%3C\/text%3E%3Ctext x='290' y='338' text-anchor='middle' font-family='Arial, sans-serif' font-size='17' font-weight='700' fill='%2300184C'%3EEvidence%3C\/text%3E%3Ctext x='531' y='338' text-anchor='middle' font-family='Arial, sans-serif' font-size='17' font-weight='700' fill='%2300184C'%3EDecision record%3C\/text%3E%3Ctext x='410' y='410' text-anchor='middle' font-family='Arial, sans-serif' font-size='20' font-weight='600' fill='%23AEC9FF'%3ETwo signals, one governed review%3C\/text%3E%3C\/svg%3E\" alt=\"rule AI risk score disagreement visual showing a rule result and AI score reviewed with explanations, evidence and a decision record\" loading=\"lazy\" width=\"820\" height=\"460\">\n    <p class=\"nx-caption\">Rule results and AI risk scores answer different questions. Explanation fields and transaction evidence help investigators understand a disagreement and record their decision. Illustrative only.<\/p>\n  <\/div>\n\n  <!-- QUICK ANSWER -->\n  <div class=\"nx-callout\">\n    <div class=\"nx-callout-title\">Quick answer<\/div>\n    <p>Neither signal should silently override the other. In FraudShield, a triggered rule stands and follows normal handling even when the AI score is low. A high AI score with no triggered rule is a complementary indication that the transaction may sit near relevant thresholds; it is not a rule breach or a confirmed fraud finding. Investigators should use the score explanations and the evidence to understand the difference, apply the prescribed route and record their decision.<\/p>\n  <\/div>\n\n  <h2 id=\"understand-the-different-jobs-of-rules-and-ai-scoring\"><span class=\"ez-toc-section\" id=\"Understand_the_Different_Jobs_of_Rules_and_AI_Scoring\"><\/span>Understand the Different Jobs of Rules and AI Scoring<span class=\"ez-toc-section-end\"><\/span><\/h2>\n  <p>In FraudShield, rules determine whether a transaction is flagged. The <a class=\"nx-inline-link\" href=\"https:\/\/nexiant.ai\/fraudshield\/docs\/guides\/ai-engine\/how-ai-risk-scoring-works\/\">AI Engine<\/a> generates a separate score from 0 to 100 when a model is enabled on the active rule plan. The score indicates how close the activity came to breaching the rules in that plan.<\/p>\n  <p>It is independent of the final rule-trigger result. It does not change merely because a rule flags the transaction, and a lower AI score does not cancel a triggered rule.<\/p>\n  <p>Equally, a high score is not a confirmed fraud finding. It is a complementary indication that an unflagged transaction may sit near one or more relevant thresholds.<\/p>\n\n  <!-- TABLE -->\n  <div class=\"nx-table-wrap\">\n    <table class=\"nx-table\">\n      <thead>\n        <tr>\n          <th>Situation<\/th>\n          <th>What it means<\/th>\n          <th>Handling<\/th>\n        <\/tr>\n      <\/thead>\n      <tbody>\n        <tr>\n          <td><strong>Rule flags, AI score low<\/strong><\/td>\n          <td>The triggered rule still requires the organisation&#8217;s normal handling. A low score is not an authorised override.<\/td>\n          <td><span class=\"nx-badge nx-badge--red\">Normal rule handling<\/span><\/td>\n        <\/tr>\n        <tr>\n          <td><strong>No rule flags, AI score high<\/strong><\/td>\n          <td>The transaction may sit just below multiple thresholds. It has not breached a rule and fraud has not been established.<\/td>\n          <td><span class=\"nx-badge nx-badge--blue\">Optional AI-based case<\/span><\/td>\n        <\/tr>\n      <\/tbody>\n    <\/table>\n  <\/div>\n\n  <h2 id=\"scenario-one-a-rule-flags-but-the-ai-score-is-low\"><span class=\"ez-toc-section\" id=\"Scenario_One_A_Rule_Flags_but_the_AI_Score_is_Low\"><\/span>Scenario One: A Rule Flags, but the AI Score is Low<span class=\"ez-toc-section-end\"><\/span><\/h2>\n  <p>Imagine a transaction that crosses one particularly important threshold but otherwise differs from the profile of transactions that tend to sit close to several rule boundaries. The rule flags it, as designed. The model&#8217;s overall score is relatively low.<\/p>\n  <p>A reviewer should not conclude that the alert can be dismissed simply because the number is modest. The triggered rule still requires the organisation&#8217;s normal handling. The reviewer can inspect why the rule was met, review relevant transaction and customer context and document the decision.<\/p>\n\n  <div class=\"nx-callout nx-callout--warning\">\n    <div class=\"nx-callout-title\">Not an override<\/div>\n    <p>The disagreement may be an opportunity to evaluate whether the rule is working as intended, but it is not an authorised override.<\/p>\n  <\/div>\n\n  <h2 id=\"scenario-two-no-rule-flags-but-the-ai-score-is-high\"><span class=\"ez-toc-section\" id=\"Scenario_Two_No_Rule_Flags_but_the_AI_Score_is_High\"><\/span>Scenario Two: No Rule Flags, but the AI Score is High<span class=\"ez-toc-section-end\"><\/span><\/h2>\n  <p>This is the use case the AI Engine is particularly intended to surface. A transaction may sit just below multiple thresholds and therefore trigger no rule, while its overall proximity to those conditions produces a higher score.<\/p>\n  <p>An organisation may choose to review such cases. FraudShield&#8217;s optional score-based auto-case creation can be configured to create a case when a transaction triggers no rules and falls within an enabled score range.<\/p>\n  <p>That path provides additional visibility; it does not mean the transaction has breached a rule or that fraudulent conduct has been established.<\/p>\n\n  <h2 id=\"use-explanations-to-investigate-the-difference\"><span class=\"ez-toc-section\" id=\"Use_Explanations_to_Investigate_the_Difference\"><\/span>Use Explanations to Investigate the Difference<span class=\"ez-toc-section-end\"><\/span><\/h2>\n  <p>FraudShield provides rule-related explanation fields for the AI score, including Impact, Influence and Direction, and identifies rules with substantial effect on the score. These help an analyst understand which underlying conditions are driving the model&#8217;s assessment.<\/p>\n  <p>A useful review sequence is:<\/p>\n\n  <!-- CONTROL STACK -->\n  <div class=\"nx-stack\">\n    <div class=\"nx-stack-layer\">\n      <div class=\"nx-stack-icon\">01<\/div>\n      <div>\n        <h4>Check the rules<\/h4>\n        <p>Check which rules triggered and which did not.<\/p>\n      <\/div>\n    <\/div>\n    <div class=\"nx-stack-layer\">\n      <div class=\"nx-stack-icon\">02<\/div>\n      <div>\n        <h4>Review the score<\/h4>\n        <p>Review the AI score and material rule-level explanations.<\/p>\n      <\/div>\n    <\/div>\n    <div class=\"nx-stack-layer\">\n      <div class=\"nx-stack-icon\">03<\/div>\n      <div>\n        <h4>Verify the facts<\/h4>\n        <p>Verify the transaction details and relevant customer context.<\/p>\n      <\/div>\n    <\/div>\n    <div class=\"nx-stack-layer\">\n      <div class=\"nx-stack-icon\">04<\/div>\n      <div>\n        <h4>Apply the route<\/h4>\n        <p>Apply the prescribed investigation or escalation route.<\/p>\n      <\/div>\n    <\/div>\n    <div class=\"nx-stack-layer\">\n      <div class=\"nx-stack-icon\">05<\/div>\n      <div>\n        <h4>Record the decision<\/h4>\n        <p>Record why the case was closed, continued or escalated.<\/p>\n      <\/div>\n    <\/div>\n  <\/div>\n\n  <p>This keeps the decision anchored in evidence rather than selecting whichever output seems more reassuring.<\/p>\n\n  <h2 id=\"what-should-monitoring-managers-learn-from-disagreement\"><span class=\"ez-toc-section\" id=\"What_Should_Monitoring_Managers_Learn_from_Disagreement\"><\/span>What Should Monitoring Managers Learn from Disagreement?<span class=\"ez-toc-section-end\"><\/span><\/h2>\n  <p>Repeated disagreements can provide a prompt for policy review. For example, a monitoring team may wish to examine whether certain rule thresholds are too broad, whether a near-threshold pattern deserves closer attention or whether the case-routing ranges produce manageable volumes.<\/p>\n  <p>Any adjustments should be tested before deployment, with a clear record of who approved them and how outcomes will be monitored. The aim is to improve coverage and investigation quality without creating an uncontrolled cascade of alerts.<\/p>\n\n  <h2 id=\"decision-ai-is-about-governed-next-steps\"><span class=\"ez-toc-section\" id=\"Decision_AI_is_About_Governed_Next_Steps\"><\/span>Decision AI is About Governed Next Steps<span class=\"ez-toc-section-end\"><\/span><\/h2>\n  <p>The broader lesson extends beyond FraudShield. Different decision signals can be complementary even when their outputs do not match. A useful Decision AI approach preserves their meaning, explains the differences and routes the result to the right action.<\/p>\n  <p>It does not collapse uncertainty into a false sense of certainty. It also does not permit a model to silently override policies that the organisation has chosen to enforce.<\/p>\n\n  <h2 id=\"review-both-signals-with-confidence\"><span class=\"ez-toc-section\" id=\"Review_Both_Signals_with_Confidence\"><\/span>Review Both Signals with Confidence<span class=\"ez-toc-section-end\"><\/span><\/h2>\n  <p>Rules and AI scoring answer different questions. See <a class=\"nx-inline-link\" href=\"https:\/\/nexiant.ai\/fraudshield\/docs\/guides\/ai-engine\/how-ai-risk-scoring-works\/\">How AI Risk Scoring Works<\/a> or <a class=\"nx-inline-link\" href=\"https:\/\/nexiant.ai\/contact-us\/\">contact Nexiant<\/a> to explore how they can complement your transaction monitoring operations.<\/p>\n\n  <hr class=\"nx-divider\">\n\n  <h2 id=\"frequently-asked-questions\"><span class=\"ez-toc-section\" id=\"Frequently_Asked_Questions\"><\/span>Frequently Asked Questions<span class=\"ez-toc-section-end\"><\/span><\/h2>\n\n  <div class=\"nx-faq\">\n    <div class=\"nx-faq-item\">\n      <button class=\"nx-faq-q\" aria-expanded=\"false\">Does a low AI score override a triggered FraudShield rule?<span class=\"nx-chevron\"><svg viewBox=\"0 0 10 6\" stroke-width=\"1.5\" stroke-linecap=\"round\" stroke-linejoin=\"round\"><path d=\"M1 1l4 4 4-4\"><\/path><\/svg><\/span><\/button>\n      <div class=\"nx-faq-a\">No. A triggered rule stands and should be handled through the usual review process.<\/div>\n    <\/div>\n    <div class=\"nx-faq-item\">\n      <button class=\"nx-faq-q\" aria-expanded=\"false\">Can a high AI score create a case without a triggered rule?<span class=\"nx-chevron\"><svg viewBox=\"0 0 10 6\" stroke-width=\"1.5\" stroke-linecap=\"round\" stroke-linejoin=\"round\"><path d=\"M1 1l4 4 4-4\"><\/path><\/svg><\/span><\/button>\n      <div class=\"nx-faq-a\">Yes, if optional AI-based case creation is enabled for the relevant score range and a model scored the transaction.<\/div>\n    <\/div>\n  <\/div>\n\n  <!-- CTA -->\n  <div class=\"nx-cta\">\n    <h3>Talk to Nexiant about rules and AI scoring<\/h3>\n    <p>Speak with our team about how rules and AI risk scoring can complement your transaction monitoring operations.<\/p>\n    <a href=\"https:\/\/nexiant.ai\/contact-us\/\">Contact Nexiant<\/a>\n  <\/div>\n\n  <p class=\"nx-disclaimer\">This article was accurate at the time of publication in November 2026 and is intended for general informational purposes only. It does not constitute legal, regulatory or compliance advice. Organisations should seek qualified professional guidance in relation to their specific obligations.<\/p>\n\n<\/div>\n\n<script>\n  document.querySelectorAll('.nx-faq-q').forEach(function(btn) {\n    btn.addEventListener('click', function() {\n      var expanded = this.getAttribute('aria-expanded') === 'true';\n      this.setAttribute('aria-expanded', !expanded);\n      this.nextElementSibling.classList.toggle('open', !expanded);\n    });\n  });\n<\/script>\n\n<script type=\"application\/ld+json\">\n{\n  \"@context\": \"https:\/\/schema.org\",\n  \"@type\": \"FAQPage\",\n  \"mainEntity\": [\n    {\n      \"@type\": \"Question\",\n      \"name\": \"Does a low AI score override a triggered FraudShield rule?\",\n      \"acceptedAnswer\": {\n        \"@type\": \"Answer\",\n        \"text\": \"No. A triggered rule stands and should be handled through the usual review process.\"\n      }\n    },\n    {\n      \"@type\": \"Question\",\n      \"name\": \"Can a high AI score create a case without a triggered rule?\",\n      \"acceptedAnswer\": {\n        \"@type\": \"Answer\",\n        \"text\": \"Yes, if optional AI-based case creation is enabled for the relevant score range and a model scored the transaction.\"\n      }\n    }\n  ]\n}\n<\/script>\n\n\n\n<p class=\"wp-block-paragraph\"><\/p>\n","protected":false},"excerpt":{"rendered":"<p>FraudShield AI Engine &nbsp;\u00b7&nbsp; November 2026 &nbsp;\u00b7&nbsp; Australia &amp; International Why different risk signals should be explained and reviewed rather than letting one overwrite the other. A rule-based monitoring system can flag a transaction while an accompanying AI score appears relatively low. The reverse is also possible: no rule triggers, yet the AI score is [&hellip;]<\/p>\n","protected":false},"author":7,"featured_media":0,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_aioseo_title":"When Transaction Rules and AI Scores Disagree | Nexiant","_aioseo_description":"What should investigators do when a rule flags but an AI score is low, or vice versa? Learn how FraudShield combines distinct signals.","om_disable_all_campaigns":false,"footnotes":""},"categories":[15,44],"tags":[83,38],"class_list":["post-947","post","type-post","status-publish","format-standard","hentry","category-risk-management","category-artificial-intelligence-ai","tag-decision-ai","tag-transaction-monitoring"],"aioseo_notices":[],"aioseo_head":"\n\t\t<!-- All in One SEO 5.0.4 - aioseo.com -->\n\t<meta name=\"description\" content=\"What should investigators do when a rule flags but an AI score is low, or vice versa? Learn how FraudShield combines distinct signals.\" \/>\n\t<meta name=\"robots\" content=\"max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n\t<meta name=\"author\" content=\"Zeeshan Rizvi\"\/>\n\t<link rel=\"canonical\" href=\"https:\/\/nexiant.ai\/resources\/blogs\/when-rules-and-ai-scores-disagree\/\" \/>\n\t\t<meta name=\"generator\" content=\"All in One SEO (AIOSEO) 5.0.4\" \/>\n\t\t<meta property=\"og:locale\" content=\"en_US\" \/>\n\t\t<meta property=\"og:site_name\" content=\"AI-Assisted Fraud Prevention and Risk Management - Nexiant\" \/>\n\t\t<meta property=\"og:type\" content=\"article\" \/>\n\t\t<meta property=\"og:title\" content=\"When Transaction Rules and AI Scores Disagree | Nexiant\" \/>\n\t\t<meta property=\"og:description\" content=\"What should investigators do when a rule flags but an AI score is low, or vice versa? Learn how FraudShield combines distinct signals.\" \/>\n\t\t<meta property=\"og:url\" content=\"https:\/\/nexiant.ai\/resources\/blogs\/when-rules-and-ai-scores-disagree\/\" \/>\n\t\t<meta property=\"og:image\" content=\"https:\/\/nexiant.ai\/resources\/blogs\/wp-content\/uploads\/2025\/11\/Nexiant-Dark-Logo-Apr.png\" \/>\n\t\t<meta property=\"og:image:secure_url\" content=\"https:\/\/nexiant.ai\/resources\/blogs\/wp-content\/uploads\/2025\/11\/Nexiant-Dark-Logo-Apr.png\" \/>\n\t\t<meta property=\"article:published_time\" content=\"2026-11-03T01:12:52+00:00\" \/>\n\t\t<meta property=\"article:modified_time\" content=\"2026-11-11T01:17:20+00:00\" \/>\n\t\t<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n\t\t<meta name=\"twitter:title\" content=\"When Transaction Rules and AI Scores Disagree | Nexiant\" \/>\n\t\t<meta name=\"twitter:description\" content=\"What should investigators do when a rule flags but an AI score is low, or vice versa? Learn how FraudShield combines distinct signals.\" \/>\n\t\t<meta name=\"twitter:image\" content=\"https:\/\/nexiant.ai\/resources\/blogs\/wp-content\/uploads\/2025\/11\/Nexiant-Dark-Logo-Apr.png\" \/>\n\t\t<script type=\"application\/ld+json\" class=\"aioseo-schema\">\n\t\t\t{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"BlogPosting\",\"@id\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/when-rules-and-ai-scores-disagree\\\/#blogposting\",\"name\":\"When Transaction Rules and AI Scores Disagree | Nexiant\",\"headline\":\"What Should Happen When a Rule and an AI Risk Score Disagree?\",\"author\":{\"@id\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/author\\\/zeeshan\\\/#author\"},\"publisher\":{\"@id\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/#organization\"},\"image\":{\"@type\":\"ImageObject\",\"url\":\"data:image\\\/svg+xml,%3Csvg xmlns='http:\\\/\\\/www.w3.org\\\/2000\\\/svg' width='820' height='460' viewBox='0 0 820 460'%3E%3Cdefs%3E%3ClinearGradient id='bg' x1='0' y1='0' x2='1' y2='1'%3E%3Cstop offset='0' stop-color='%2300184C'\\\/%3E%3Cstop offset='1' stop-color='%23073EA1'\\\/%3E%3C\\\/linearGradient%3E%3C\\\/defs%3E%3Crect width='820' height='460' rx='22' fill='url(%23bg)'\\\/%3E%3Ctext x='410' y='70' text-anchor='middle' font-family='Arial, sans-serif' font-size='30' font-weight='700' fill='%23ffffff'%3EWhen Rules and AI Scores Disagree%3C\\\/text%3E%3Crect x='86' y='132' width='165' height='92' rx='18' fill='%23ffffff' opacity='0.92'\\\/%3E%3Crect x='328' y='132' width='165' height='92' rx='18' fill='%23ffffff' opacity='0.88'\\\/%3E%3Crect x='570' y='132' width='165' height='92' rx='18' fill='%23ffffff' opacity='0.84'\\\/%3E%3Crect x='207' y='272' width='165' height='92' rx='18' fill='%23ffffff' opacity='0.86'\\\/%3E%3Crect x='449' y='272' width='165' height='92' rx='18' fill='%23ffffff' opacity='0.9'\\\/%3E%3Ccircle cx='168' cy='162' r='16' fill='%23073EA1'\\\/%3E%3Ccircle cx='410' cy='162' r='16' fill='%23073EA1'\\\/%3E%3Ccircle cx='652' cy='162' r='16' fill='%23073EA1'\\\/%3E%3Ccircle cx='290' cy='302' r='16' fill='%23073EA1'\\\/%3E%3Ccircle cx='531' cy='302' r='16' fill='%23073EA1'\\\/%3E%3Ctext x='168' y='198' text-anchor='middle' font-family='Arial, sans-serif' font-size='17' font-weight='700' fill='%2300184C'%3ERule result%3C\\\/text%3E%3Ctext x='410' y='198' text-anchor='middle' font-family='Arial, sans-serif' font-size='17' font-weight='700' fill='%2300184C'%3EAI score%3C\\\/text%3E%3Ctext x='652' y='198' text-anchor='middle' font-family='Arial, sans-serif' font-size='17' font-weight='700' fill='%2300184C'%3EExplanation%3C\\\/text%3E%3Ctext x='290' y='338' text-anchor='middle' font-family='Arial, sans-serif' font-size='17' font-weight='700' fill='%2300184C'%3EEvidence%3C\\\/text%3E%3Ctext x='531' y='338' text-anchor='middle' font-family='Arial, sans-serif' font-size='17' font-weight='700' fill='%2300184C'%3EDecision record%3C\\\/text%3E%3Ctext x='410' y='410' text-anchor='middle' font-family='Arial, sans-serif' font-size='20' font-weight='600' fill='%23AEC9FF'%3ETwo signals, one governed review%3C\\\/text%3E%3C\\\/svg%3E\",\"@id\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/when-rules-and-ai-scores-disagree\\\/#articleImage\"},\"datePublished\":\"2026-11-03T12:12:52+11:00\",\"dateModified\":\"2026-11-11T12:17:20+11:00\",\"inLanguage\":\"en-AU\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/when-rules-and-ai-scores-disagree\\\/#webpage\"},\"isPartOf\":{\"@id\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/when-rules-and-ai-scores-disagree\\\/#webpage\"},\"articleSection\":\"Risk Management, Artificial Intelligence (AI), Decision AI, Transaction Monitoring\"},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/when-rules-and-ai-scores-disagree\\\/#breadcrumblist\",\"itemListElement\":[{\"@type\":\"ListItem\",\"@id\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/#listItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/\",\"nextItem\":{\"@type\":\"ListItem\",\"@id\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/category\\\/risk-management\\\/#listItem\",\"name\":\"Risk Management\"}},{\"@type\":\"ListItem\",\"@id\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/category\\\/risk-management\\\/#listItem\",\"position\":2,\"name\":\"Risk Management\",\"item\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/category\\\/risk-management\\\/\",\"nextItem\":{\"@type\":\"ListItem\",\"@id\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/when-rules-and-ai-scores-disagree\\\/#listItem\",\"name\":\"What Should Happen When a Rule and an AI Risk Score Disagree?\"},\"previousItem\":{\"@type\":\"ListItem\",\"@id\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/#listItem\",\"name\":\"Home\"}},{\"@type\":\"ListItem\",\"@id\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/when-rules-and-ai-scores-disagree\\\/#listItem\",\"position\":3,\"name\":\"What Should Happen When a Rule and an AI Risk Score Disagree?\",\"previousItem\":{\"@type\":\"ListItem\",\"@id\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/category\\\/risk-management\\\/#listItem\",\"name\":\"Risk Management\"}}]},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/#organization\",\"name\":\"AI-Assisted Fraud Prevention and Risk Management\",\"description\":\"Nexiant\",\"url\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/\",\"logo\":{\"@type\":\"ImageObject\",\"url\":\"https:\\\/\\\/wp-nexiant.nexiant.ai\\\/wp-content\\\/uploads\\\/2025\\\/11\\\/Nexiant-Dark-Logo-Apr.png\",\"@id\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/when-rules-and-ai-scores-disagree\\\/#organizationLogo\"},\"image\":{\"@id\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/when-rules-and-ai-scores-disagree\\\/#organizationLogo\"},\"sameAs\":[\"https:\\\/\\\/www.linkedin.com\\\/company\\\/nexiantai\"]},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/author\\\/zeeshan\\\/#author\",\"url\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/author\\\/zeeshan\\\/\",\"name\":\"Zeeshan Rizvi\",\"image\":{\"@type\":\"ImageObject\",\"@id\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/when-rules-and-ai-scores-disagree\\\/#authorImage\",\"url\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/5060e430addd1087c0299e5ea76cd12ea794a85047ffecd5cfbbf24f4c1fbb2f?s=96&d=mm&r=g\",\"width\":96,\"height\":96,\"caption\":\"Zeeshan Rizvi\"}},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/when-rules-and-ai-scores-disagree\\\/#webpage\",\"url\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/when-rules-and-ai-scores-disagree\\\/\",\"name\":\"When Transaction Rules and AI Scores Disagree | Nexiant\",\"description\":\"What should investigators do when a rule flags but an AI score is low, or vice versa? Learn how FraudShield combines distinct signals.\",\"inLanguage\":\"en-AU\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/#website\"},\"breadcrumb\":{\"@id\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/when-rules-and-ai-scores-disagree\\\/#breadcrumblist\"},\"author\":{\"@id\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/author\\\/zeeshan\\\/#author\"},\"creator\":{\"@id\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/author\\\/zeeshan\\\/#author\"},\"datePublished\":\"2026-11-03T12:12:52+11:00\",\"dateModified\":\"2026-11-11T12:17:20+11:00\"},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/#website\",\"url\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/\",\"name\":\"AI-Assisted Fraud Prevention and Risk Management\",\"description\":\"Nexiant\",\"inLanguage\":\"en-AU\",\"publisher\":{\"@id\":\"https:\\\/\\\/nexiant.ai\\\/resources\\\/blogs\\\/#organization\"}}]}\n\t\t<\/script>\n\t\t<!-- All in One SEO -->\n\n","aioseo_head_json":{"title":"When Transaction Rules and AI Scores Disagree | Nexiant","description":"What should investigators do when a rule flags but an AI score is low, or vice versa? Learn how FraudShield combines distinct signals.","canonical_url":"https:\/\/nexiant.ai\/resources\/blogs\/when-rules-and-ai-scores-disagree\/","robots":"max-snippet:-1, max-image-preview:large, max-video-preview:-1","keywords":"","webmasterTools":{"miscellaneous":""},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"BlogPosting","@id":"https:\/\/nexiant.ai\/resources\/blogs\/when-rules-and-ai-scores-disagree\/#blogposting","name":"When Transaction Rules and AI Scores Disagree | Nexiant","headline":"What Should Happen When a Rule and an AI Risk Score Disagree?","author":{"@id":"https:\/\/nexiant.ai\/resources\/blogs\/author\/zeeshan\/#author"},"publisher":{"@id":"https:\/\/nexiant.ai\/resources\/blogs\/#organization"},"image":{"@type":"ImageObject","url":"data:image\/svg+xml,%3Csvg xmlns='http:\/\/www.w3.org\/2000\/svg' width='820' height='460' viewBox='0 0 820 460'%3E%3Cdefs%3E%3ClinearGradient id='bg' x1='0' y1='0' x2='1' y2='1'%3E%3Cstop offset='0' stop-color='%2300184C'\/%3E%3Cstop offset='1' stop-color='%23073EA1'\/%3E%3C\/linearGradient%3E%3C\/defs%3E%3Crect width='820' height='460' rx='22' fill='url(%23bg)'\/%3E%3Ctext x='410' y='70' text-anchor='middle' font-family='Arial, sans-serif' font-size='30' font-weight='700' fill='%23ffffff'%3EWhen Rules and AI Scores Disagree%3C\/text%3E%3Crect x='86' y='132' width='165' height='92' rx='18' fill='%23ffffff' opacity='0.92'\/%3E%3Crect x='328' y='132' width='165' height='92' rx='18' fill='%23ffffff' opacity='0.88'\/%3E%3Crect x='570' y='132' width='165' height='92' rx='18' fill='%23ffffff' opacity='0.84'\/%3E%3Crect x='207' y='272' width='165' height='92' rx='18' fill='%23ffffff' opacity='0.86'\/%3E%3Crect x='449' y='272' width='165' height='92' rx='18' fill='%23ffffff' opacity='0.9'\/%3E%3Ccircle cx='168' cy='162' r='16' fill='%23073EA1'\/%3E%3Ccircle cx='410' cy='162' r='16' fill='%23073EA1'\/%3E%3Ccircle cx='652' cy='162' r='16' fill='%23073EA1'\/%3E%3Ccircle cx='290' cy='302' r='16' fill='%23073EA1'\/%3E%3Ccircle cx='531' cy='302' r='16' fill='%23073EA1'\/%3E%3Ctext x='168' y='198' text-anchor='middle' font-family='Arial, sans-serif' font-size='17' font-weight='700' fill='%2300184C'%3ERule result%3C\/text%3E%3Ctext x='410' y='198' text-anchor='middle' font-family='Arial, sans-serif' font-size='17' font-weight='700' fill='%2300184C'%3EAI score%3C\/text%3E%3Ctext x='652' y='198' text-anchor='middle' font-family='Arial, sans-serif' font-size='17' font-weight='700' fill='%2300184C'%3EExplanation%3C\/text%3E%3Ctext x='290' y='338' text-anchor='middle' font-family='Arial, sans-serif' font-size='17' font-weight='700' fill='%2300184C'%3EEvidence%3C\/text%3E%3Ctext x='531' y='338' text-anchor='middle' font-family='Arial, sans-serif' font-size='17' font-weight='700' fill='%2300184C'%3EDecision record%3C\/text%3E%3Ctext x='410' y='410' text-anchor='middle' font-family='Arial, sans-serif' font-size='20' font-weight='600' fill='%23AEC9FF'%3ETwo signals, one governed review%3C\/text%3E%3C\/svg%3E","@id":"https:\/\/nexiant.ai\/resources\/blogs\/when-rules-and-ai-scores-disagree\/#articleImage"},"datePublished":"2026-11-03T12:12:52+11:00","dateModified":"2026-11-11T12:17:20+11:00","inLanguage":"en-AU","mainEntityOfPage":{"@id":"https:\/\/nexiant.ai\/resources\/blogs\/when-rules-and-ai-scores-disagree\/#webpage"},"isPartOf":{"@id":"https:\/\/nexiant.ai\/resources\/blogs\/when-rules-and-ai-scores-disagree\/#webpage"},"articleSection":"Risk Management, Artificial Intelligence (AI), Decision AI, Transaction Monitoring"},{"@type":"BreadcrumbList","@id":"https:\/\/nexiant.ai\/resources\/blogs\/when-rules-and-ai-scores-disagree\/#breadcrumblist","itemListElement":[{"@type":"ListItem","@id":"https:\/\/nexiant.ai\/resources\/blogs\/#listItem","position":1,"name":"Home","item":"https:\/\/nexiant.ai\/resources\/blogs\/","nextItem":{"@type":"ListItem","@id":"https:\/\/nexiant.ai\/resources\/blogs\/category\/risk-management\/#listItem","name":"Risk Management"}},{"@type":"ListItem","@id":"https:\/\/nexiant.ai\/resources\/blogs\/category\/risk-management\/#listItem","position":2,"name":"Risk Management","item":"https:\/\/nexiant.ai\/resources\/blogs\/category\/risk-management\/","nextItem":{"@type":"ListItem","@id":"https:\/\/nexiant.ai\/resources\/blogs\/when-rules-and-ai-scores-disagree\/#listItem","name":"What Should Happen When a Rule and an AI Risk Score Disagree?"},"previousItem":{"@type":"ListItem","@id":"https:\/\/nexiant.ai\/resources\/blogs\/#listItem","name":"Home"}},{"@type":"ListItem","@id":"https:\/\/nexiant.ai\/resources\/blogs\/when-rules-and-ai-scores-disagree\/#listItem","position":3,"name":"What Should Happen When a Rule and an AI Risk Score Disagree?","previousItem":{"@type":"ListItem","@id":"https:\/\/nexiant.ai\/resources\/blogs\/category\/risk-management\/#listItem","name":"Risk Management"}}]},{"@type":"Organization","@id":"https:\/\/nexiant.ai\/resources\/blogs\/#organization","name":"AI-Assisted Fraud Prevention and Risk Management","description":"Nexiant","url":"https:\/\/nexiant.ai\/resources\/blogs\/","logo":{"@type":"ImageObject","url":"https:\/\/wp-nexiant.nexiant.ai\/wp-content\/uploads\/2025\/11\/Nexiant-Dark-Logo-Apr.png","@id":"https:\/\/nexiant.ai\/resources\/blogs\/when-rules-and-ai-scores-disagree\/#organizationLogo"},"image":{"@id":"https:\/\/nexiant.ai\/resources\/blogs\/when-rules-and-ai-scores-disagree\/#organizationLogo"},"sameAs":["https:\/\/www.linkedin.com\/company\/nexiantai"]},{"@type":"Person","@id":"https:\/\/nexiant.ai\/resources\/blogs\/author\/zeeshan\/#author","url":"https:\/\/nexiant.ai\/resources\/blogs\/author\/zeeshan\/","name":"Zeeshan Rizvi","image":{"@type":"ImageObject","@id":"https:\/\/nexiant.ai\/resources\/blogs\/when-rules-and-ai-scores-disagree\/#authorImage","url":"https:\/\/secure.gravatar.com\/avatar\/5060e430addd1087c0299e5ea76cd12ea794a85047ffecd5cfbbf24f4c1fbb2f?s=96&d=mm&r=g","width":96,"height":96,"caption":"Zeeshan Rizvi"}},{"@type":"WebPage","@id":"https:\/\/nexiant.ai\/resources\/blogs\/when-rules-and-ai-scores-disagree\/#webpage","url":"https:\/\/nexiant.ai\/resources\/blogs\/when-rules-and-ai-scores-disagree\/","name":"When Transaction Rules and AI Scores Disagree | Nexiant","description":"What should investigators do when a rule flags but an AI score is low, or vice versa? Learn how FraudShield combines distinct signals.","inLanguage":"en-AU","isPartOf":{"@id":"https:\/\/nexiant.ai\/resources\/blogs\/#website"},"breadcrumb":{"@id":"https:\/\/nexiant.ai\/resources\/blogs\/when-rules-and-ai-scores-disagree\/#breadcrumblist"},"author":{"@id":"https:\/\/nexiant.ai\/resources\/blogs\/author\/zeeshan\/#author"},"creator":{"@id":"https:\/\/nexiant.ai\/resources\/blogs\/author\/zeeshan\/#author"},"datePublished":"2026-11-03T12:12:52+11:00","dateModified":"2026-11-11T12:17:20+11:00"},{"@type":"WebSite","@id":"https:\/\/nexiant.ai\/resources\/blogs\/#website","url":"https:\/\/nexiant.ai\/resources\/blogs\/","name":"AI-Assisted Fraud Prevention and Risk Management","description":"Nexiant","inLanguage":"en-AU","publisher":{"@id":"https:\/\/nexiant.ai\/resources\/blogs\/#organization"}}]},"og:locale":"en_US","og:site_name":"AI-Assisted Fraud Prevention and Risk Management - Nexiant","og:type":"article","og:title":"When Transaction Rules and AI Scores Disagree | Nexiant","og:description":"What should investigators do when a rule flags but an AI score is low, or vice versa? Learn how FraudShield combines distinct signals.","og:url":"https:\/\/nexiant.ai\/resources\/blogs\/when-rules-and-ai-scores-disagree\/","og:image":"https:\/\/nexiant.ai\/resources\/blogs\/wp-content\/uploads\/2025\/11\/Nexiant-Dark-Logo-Apr.png","og:image:secure_url":"https:\/\/nexiant.ai\/resources\/blogs\/wp-content\/uploads\/2025\/11\/Nexiant-Dark-Logo-Apr.png","article:published_time":"2026-11-03T01:12:52+00:00","article:modified_time":"2026-11-11T01:17:20+00:00","twitter:card":"summary_large_image","twitter:title":"When Transaction Rules and AI Scores Disagree | Nexiant","twitter:description":"What should investigators do when a rule flags but an AI score is low, or vice versa? Learn how FraudShield combines distinct signals.","twitter:image":"https:\/\/nexiant.ai\/resources\/blogs\/wp-content\/uploads\/2025\/11\/Nexiant-Dark-Logo-Apr.png"},"aioseo_meta_data":{"post_id":"947","title":"When Transaction Rules and AI Scores Disagree | Nexiant","description":"What should investigators do when a rule flags but an AI score is low, or vice versa? Learn how FraudShield combines distinct signals.","keywords":null,"keyphrases":{"focus":{"keyphrase":"rule AI risk score disagreement","score":0,"analysis":[]},"additional":[]},"primary_term":null,"canonical_url":null,"og_title":null,"og_description":null,"og_object_type":"default","og_image_type":"default","og_image_url":null,"og_image_width":null,"og_image_height":null,"og_image_custom_url":null,"og_image_custom_fields":null,"og_video":"","og_custom_url":null,"og_article_section":null,"og_article_tags":null,"twitter_use_og":false,"twitter_card":"default","twitter_image_type":"default","twitter_image_url":null,"twitter_image_custom_url":null,"twitter_image_custom_fields":null,"twitter_title":null,"twitter_description":null,"schema":{"blockGraphs":[],"customGraphs":[],"default":{"data":{"Article":[],"Course":[],"Dataset":[],"FAQPage":[],"Movie":[],"Person":[],"Product":[],"ProductReview":[],"Car":[],"Recipe":[],"Service":[],"SoftwareApplication":[],"WebPage":[]},"graphName":"BlogPosting","isEnabled":true},"graphs":[]},"schema_type":"default","schema_type_options":null,"pillar_content":false,"robots_default":true,"robots_noindex":false,"robots_noarchive":false,"robots_nosnippet":false,"robots_nofollow":false,"robots_noimageindex":false,"robots_noodp":false,"robots_notranslate":false,"robots_max_snippet":"-1","robots_max_videopreview":"-1","robots_max_imagepreview":"large","priority":null,"frequency":"default","local_seo":null,"breadcrumb_settings":null,"limit_modified_date":false,"ai":{"faqs":[],"keyPoints":[],"schemas":[],"titles":[],"descriptions":[],"socialPosts":{"email":{"subject":"","preview":"","content":""},"linkedin":[],"twitter":[],"facebook":[],"instagram":[]}},"created":"2026-11-11 01:17:20","updated":"2026-11-11 01:55:09","focus_keyword":"rule AI risk score disagreement","additional_keywords":null,"truseo_locale":null,"seo_analyzer_scan_date":null},"aioseo_breadcrumb":"<div class=\"aioseo-breadcrumbs\"><span class=\"aioseo-breadcrumb\">\n\t\t\t<a href=\"https:\/\/nexiant.ai\/resources\/blogs\/\" title=\"Home\">Home<\/a>\n\t\t<\/span><span class=\"aioseo-breadcrumb-separator\">\u00bb<\/span><span class=\"aioseo-breadcrumb\">\n\t\t\t<a href=\"https:\/\/nexiant.ai\/resources\/blogs\/category\/risk-management\/\" title=\"Risk Management\">Risk Management<\/a>\n\t\t<\/span><span class=\"aioseo-breadcrumb-separator\">\u00bb<\/span><span class=\"aioseo-breadcrumb\">\n\t\t\tWhat Should Happen When a Rule and an AI Risk Score Disagree?\n\t\t<\/span><\/div>","aioseo_breadcrumb_json":[{"label":"Home","link":"https:\/\/nexiant.ai\/resources\/blogs\/"},{"label":"Risk Management","link":"https:\/\/nexiant.ai\/resources\/blogs\/category\/risk-management\/"},{"label":"What Should Happen When a Rule and an AI Risk Score Disagree?","link":"https:\/\/nexiant.ai\/resources\/blogs\/when-rules-and-ai-scores-disagree\/"}],"_links":{"self":[{"href":"https:\/\/nexiant.ai\/resources\/blogs\/wp-json\/wp\/v2\/posts\/947","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/nexiant.ai\/resources\/blogs\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/nexiant.ai\/resources\/blogs\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/nexiant.ai\/resources\/blogs\/wp-json\/wp\/v2\/users\/7"}],"replies":[{"embeddable":true,"href":"https:\/\/nexiant.ai\/resources\/blogs\/wp-json\/wp\/v2\/comments?post=947"}],"version-history":[{"count":1,"href":"https:\/\/nexiant.ai\/resources\/blogs\/wp-json\/wp\/v2\/posts\/947\/revisions"}],"predecessor-version":[{"id":948,"href":"https:\/\/nexiant.ai\/resources\/blogs\/wp-json\/wp\/v2\/posts\/947\/revisions\/948"}],"wp:attachment":[{"href":"https:\/\/nexiant.ai\/resources\/blogs\/wp-json\/wp\/v2\/media?parent=947"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/nexiant.ai\/resources\/blogs\/wp-json\/wp\/v2\/categories?post=947"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/nexiant.ai\/resources\/blogs\/wp-json\/wp\/v2\/tags?post=947"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}