[{"data":1,"prerenderedAt":246},["ShallowReactive",2],{"/en/workflows/signal-reliability-navigator":3},{"id":4,"slug":5,"locale":6,"translationGroupId":7,"localeSwitchApproved":8,"title":9,"description":10,"documentationMarkdown":11,"workflowJson":12,"category":229,"tags":230,"integrations":231,"difficulty":234,"author":235,"verified":34,"featured":34,"date":236,"modified":236,"icon":7,"imageSrc":7,"path":237,"alternates":238,"seo":239},"955685aa-c2ab-4efe-9ab9-940e849228dc","signal-reliability-navigator","en",null,true,"Signal Reliability Navigator","A conversational workflow that helps teams separate trustworthy branch signals from polished noise, catch dirty data early, and choose when to rely on automation vs human judgment.","## How it works\nThis workflow runs as a practical “signal coach” inside Calypso. It takes messy branch numbers, conversation notes, and event snippets and turns them into decision-ready guidance—without pretending every metric deserves a seat at the table.\n\nOperators (or leaders) pick the situation they’re in from a menu. The workflow first applies your Knowledge Base policy (so answers stay consistent with your internal definitions), then routes users to sharp, decision-shaped checks: what to trust, what’s likely dirty, what teams misread when comparing branches, and when automation should *not* be left unattended.\n\n## Key features\n- Uses a Knowledge Base policy first to keep definitions and guardrails consistent\n- Button-based menu for fast routing (no “type a paragraph and hope”)\n- Covers six common decision scenarios: trust, dirt, automation, messy evidence, comparisons, and culture\n- Loops back to the menu after each answer so users can run multiple checks in one session\n- Optional handoff to a human via routing for high-stakes or ambiguous cases\n\n## Step-by-step\n1. **Trigger:** A user starts the workflow (Input).\n2. **Apply guardrails:** Calypso enforces the **Knowledge Base policy** to keep advice aligned with your internal standards.\n3. **Choose a scenario:** The workflow sends an **interactive button menu** with the most common “this meeting is about to go wrong” situations.\n4. **Route by selection:** A chain of **IF** checks matches the button clicked.\n5. **Deliver decision guidance:** A **text message** provides practical checks and next steps for that scenario.\n6. **Keep momentum:** The workflow returns the user to the main menu to run another check.\n7. **Escalate when needed:** If the user chooses “Talk to a human” (or nothing matches), the workflow sends a **fallback/handoff** message and routes to the configured department.\n\n## Setup requirements\n- **Calypso Knowledge Base**: Recommended (the workflow is designed to use it). No external credentials required.\n- **Routing department**: Create/identify a department (e.g., “Decision Support”) for handoffs if you want the human escalation path.",{"id":13,"teamId":14,"name":9,"version":15,"workflowVersion":16,"nodes":17,"connections":191,"routingEnabled":8,"active":34},"wf_signal_reliability_navigator_v1","calypso-public-library","1.0.0",1,[18,35,41,53,86,95,104,110,117,123,130,136,143,149,156,162,168,174,185],{"id":19,"name":20,"type":21,"typeVersion":16,"position":22,"parameters":24,"category":33,"deletable":34,"connectable":34},"node_flow_cfg","Workflow settings","flow-configs",[23,23],80,{"name":9,"description":25,"tags":26,"triggerType":32},"Menu-driven coaching to separate trustworthy branch signals from polished noise, catch dirty data early, and decide when automation needs human judgment.",[27,28,29,30,31],"decision-making","signal-quality","branch-metrics","data-hygiene","automation-guardrails","input","policy",false,{"id":36,"name":37,"type":32,"typeVersion":16,"position":38,"parameters":40,"category":32,"deletable":34,"connectable":8},"node_input","Start",[23,39],220,{},{"id":42,"name":43,"type":44,"typeVersion":16,"position":45,"parameters":47,"category":52,"deletable":8,"connectable":8},"node_kb_policy","Knowledge base guardrails","knowledge-base-policy",[46,39],320,{"enabled":8,"fallbackToRouting":8,"sticky":8,"stickyMode":48,"activationOpener":49,"personalization":51},"ai_sticky_release",{"enabled":8,"instruction":50},"Use the Knowledge Base definitions for branch metrics, attribution rules, and data quality standards. Keep advice practical and decision-shaped. If confidence is low or data is ambiguous, recommend handoff.",{"useContactName":34},"response",{"id":54,"name":55,"type":56,"typeVersion":16,"position":57,"parameters":59,"category":52,"deletable":8,"connectable":8},"node_menu","Decision menu","interactive-message",[58,39],560,{"messageType":60,"headerText":9,"bodyText":61,"footerText":62,"sectionTitle":63,"buttons":64,"ctaDisplayText":63,"ctaUrl":63},"button","Pick what you’re trying to decide. I’ll help you separate trustworthy signal from polished noise (before a confident meeting makes a wrong call).","Choose one option below.","",[65,68,71,74,77,80,83],{"id":66,"title":67},"trust_numbers","Trust numbers",{"id":69,"title":70},"spot_dirty_signal","Dirty signal check",{"id":72,"title":73},"automation_vs_judgment","Auto vs human",{"id":75,"title":76},"messy_evidence","Messy to insight",{"id":78,"title":79},"comparison_traps","Compare branches",{"id":81,"title":82},"signal_culture","Signal culture",{"id":84,"title":85},"talk_human","Talk to human",{"id":87,"name":88,"type":89,"typeVersion":16,"position":90,"parameters":92,"category":94,"deletable":8,"connectable":8},"node_if_trust","If: trust_numbers","if",[91,39],800,{"buttonId":66,"operator":93},"equals","routing",{"id":96,"name":97,"type":98,"typeVersion":16,"position":99,"parameters":102,"category":52,"deletable":8,"connectable":8},"node_text_trust","Trustworthy numbers checklist","text-message",[100,101],1040,120,{"text":103},"Here’s the quick test for branch numbers you can *actually* trust:\n\n1) **Definition stability:** Did the definition change this quarter (or “quietly” last week)? If yes, treat trends as storytelling, not evidence.\n2) **Coverage & missingness:** Ask “What % of reality does this number see?” A metric with 85% coverage is fine—*as long as you don’t pretend it’s 100%.*\n3) **Lag & revision:** If it gets revised later, don’t use the first print for irreversible decisions.\n4) **Incentives:** If someone’s comp or reputation moves with it, expect polishing.\n5) **Cross-check:** One independent cross-check beats three extra decimals.\n\nRule of thumb: **precision is not accuracy wearing a nicer suit.**",{"id":105,"name":106,"type":89,"typeVersion":16,"position":107,"parameters":109,"category":94,"deletable":8,"connectable":8},"node_if_dirty","If: spot_dirty_signal",[91,108],340,{"buttonId":69,"operator":93},{"id":111,"name":112,"type":98,"typeVersion":16,"position":113,"parameters":115,"category":52,"deletable":8,"connectable":8},"node_text_dirty","Dirty signal early warnings",[100,114],260,{"text":116},"Dirty data usually shows up as *clean-looking* charts. Here are the tells before the meeting goes off the rails:\n\n- **Too-smooth jumps:** Step-changes at month-end often mean process or tagging changed, not customer behavior.\n- **Perfect symmetry:** Every branch improves “a bit” at the same time? That’s a reporting artifact until proven otherwise.\n- **Single-field dependence:** If one dropdown choice drives the whole metric, expect drift.\n- **Funnel math breaks:** If downstream counts exceed upstream, your event definitions are leaking.\n- **Narrative arrives early:** When the story shows up before the evidence, people stop noticing anomalies.\n\nFast move: ask “**What would we expect to see if this were wrong?**” Then look for *that*.",{"id":118,"name":119,"type":89,"typeVersion":16,"position":120,"parameters":122,"category":94,"deletable":8,"connectable":8},"node_if_auto","If: automation_vs_judgment",[91,121],460,{"buttonId":72,"operator":93},{"id":124,"name":125,"type":98,"typeVersion":16,"position":126,"parameters":128,"category":52,"deletable":8,"connectable":8},"node_text_auto","When to trust automation",[100,127],400,{"text":129},"Automation is great at *repeatable* decisions. It’s terrible at **novelty, incentives, and edge cases**.\n\nTrust automation when:\n- Inputs are **stable and well-defined**\n- Errors are **cheap and reversible**\n- You can measure outcomes with a tight feedback loop\n\nRequire human judgment when:\n- The metric can be **gamed** (and someone benefits)\n- The environment just changed (pricing, policy, staffing, seasonality)\n- The decision is **one-way** (customer harm, compliance risk, big spend)\n\nPractical guardrail: keep automation in “**recommend**” mode until you’ve seen it behave through at least one weird month.",{"id":131,"name":132,"type":89,"typeVersion":16,"position":133,"parameters":135,"category":94,"deletable":8,"connectable":8},"node_if_messy","If: messy_evidence",[91,134],580,{"buttonId":75,"operator":93},{"id":137,"name":138,"type":98,"typeVersion":16,"position":139,"parameters":141,"category":52,"deletable":8,"connectable":8},"node_text_messy","Messy evidence to usable insight",[100,140],540,{"text":142},"Cleaning is good—until it cleans away the truth.\n\nUse this sequence:\n1) **Preserve raw reality:** Keep the original notes/events available. Don’t overwrite; annotate.\n2) **Separate “unknown” from “no”:** Most bad decisions start when missing data gets treated as zero.\n3) **Normalize lightly:** Standardize only what you must (dates, branch IDs, basic categories). Over-normalizing can erase useful weirdness.\n4) **Tag uncertainty:** Add confidence bands in plain language: “high / medium / low” beats fake precision.\n5) **Decide with thresholds:** Define what would change your action (e.g., “If X is above 5% for 2 weeks, we intervene”).\n\nIf the insight disappears when you stop over-cleaning, it wasn’t an insight—it was a formatting trick.",{"id":144,"name":145,"type":89,"typeVersion":16,"position":146,"parameters":148,"category":94,"deletable":8,"connectable":8},"node_if_compare","If: comparison_traps",[91,147],700,{"buttonId":78,"operator":93},{"id":150,"name":151,"type":98,"typeVersion":16,"position":152,"parameters":154,"category":52,"deletable":8,"connectable":8},"node_text_compare","Branch comparison traps",[100,153],680,{"text":155},"What teams misread first when comparing branches, conversations, and attribution:\n\n- **Different denominators:** “Conversion” across branches is meaningless if footfall, lead quality, or eligibility differs.\n- **Selection effects:** The branch that gets the hardest cases will look “worse” unless you stratify.\n- **Attribution vanity:** Last-touch makes the nearest conversation look like the cause. Correlation loves a tight timestamp.\n- **Process drift:** One branch follows the script; another improvises. Your metric becomes a proxy for compliance, not customer value.\n\nMinimum safe comparison:\n1) Match cohorts (who/what/when)\n2) Compare distributions, not just averages\n3) Look for one independent cross-check (complaints, churn, repeat visits)\n\nIf you can’t explain *why* branches differ, you’re probably measuring their paperwork.",{"id":157,"name":158,"type":89,"typeVersion":16,"position":159,"parameters":161,"category":94,"deletable":8,"connectable":8},"node_if_culture","If: signal_culture",[91,160],820,{"buttonId":81,"operator":93},{"id":163,"name":164,"type":98,"typeVersion":16,"position":165,"parameters":166,"category":52,"deletable":8,"connectable":8},"node_text_culture","Build a signal culture",[100,160],{"text":167},"A healthy signal culture makes decisions happen—slides are optional.\n\nDo these and you’ll feel the difference fast:\n- **Name the decision first:** Every metric request must answer “What will we do differently?”\n- **One owner per metric:** Not for politics—so someone is accountable for definitions and breakages.\n- **Pre-mortem the dashboard:** Ask “How will this mislead us?” before it does.\n- **Celebrate anomaly reports:** Reward the person who finds the flaw *before* the rollout.\n- **Keep a ‘metric change log’:** Quiet definition changes are where trust goes to die.\n\nWit-sized truth: if your weekly meeting needs a legend to interpret the numbers, the numbers aren’t ready.",{"id":169,"name":170,"type":89,"typeVersion":16,"position":171,"parameters":173,"category":94,"deletable":8,"connectable":8},"node_if_human","If: talk_human",[91,172],940,{"buttonId":84,"operator":93},{"id":175,"name":176,"type":177,"typeVersion":16,"position":178,"parameters":180,"category":184,"deletable":8,"connectable":8},"node_fallback_handoff","Handoff to Decision Support","fallback",[100,179],960,{"handoffMessage":181,"departmentId":182,"departmentName":183},"Got it. This sounds like one of those cases where context matters more than confidence. I’m handing this to Decision Support so a human can sanity-check the signal with you.","decision-support","Decision Support","terminal",{"id":186,"name":187,"type":177,"typeVersion":16,"position":188,"parameters":189,"category":184,"deletable":8,"connectable":8},"node_fallback_default","Fallback (no selection)",[100,100],{"handoffMessage":190,"departmentId":182,"departmentName":183},"I didn’t catch a selection. If this is high-stakes or the data feels ‘too clean,’ I can route you to Decision Support for a human check.",[192,195,197,199,202,205,207,209,211,213,215,217,219,221,223,225,227],{"id":193,"source":36,"target":42,"sourceHandle":63,"targetHandle":63,"type":194},"conn_input_to_kb","default",{"id":196,"source":42,"target":54,"sourceHandle":63,"targetHandle":63,"type":194},"conn_kb_to_menu",{"id":198,"source":54,"target":87,"sourceHandle":63,"targetHandle":63,"type":194},"conn_menu_to_if_trust",{"id":200,"source":87,"target":96,"sourceHandle":201,"targetHandle":63,"type":194},"conn_if_trust_true_to_text","true",{"id":203,"source":87,"target":105,"sourceHandle":204,"targetHandle":63,"type":194},"conn_if_trust_false_to_if_dirty","false",{"id":206,"source":105,"target":111,"sourceHandle":201,"targetHandle":63,"type":194},"conn_if_dirty_true_to_text",{"id":208,"source":105,"target":118,"sourceHandle":204,"targetHandle":63,"type":194},"conn_if_dirty_false_to_if_auto",{"id":210,"source":118,"target":124,"sourceHandle":201,"targetHandle":63,"type":194},"conn_if_auto_true_to_text",{"id":212,"source":118,"target":131,"sourceHandle":204,"targetHandle":63,"type":194},"conn_if_auto_false_to_if_messy",{"id":214,"source":131,"target":137,"sourceHandle":201,"targetHandle":63,"type":194},"conn_if_messy_true_to_text",{"id":216,"source":131,"target":144,"sourceHandle":204,"targetHandle":63,"type":194},"conn_if_messy_false_to_if_compare",{"id":218,"source":144,"target":150,"sourceHandle":201,"targetHandle":63,"type":194},"conn_if_compare_true_to_text",{"id":220,"source":144,"target":157,"sourceHandle":204,"targetHandle":63,"type":194},"conn_if_compare_false_to_if_culture",{"id":222,"source":157,"target":163,"sourceHandle":201,"targetHandle":63,"type":194},"conn_if_culture_true_to_text",{"id":224,"source":157,"target":169,"sourceHandle":204,"targetHandle":63,"type":194},"conn_if_culture_false_to_if_human",{"id":226,"source":169,"target":175,"sourceHandle":201,"targetHandle":63,"type":194},"conn_if_human_true_to_handoff",{"id":228,"source":169,"target":186,"sourceHandle":204,"targetHandle":63,"type":194},"conn_if_human_false_to_default_fallback","automation",[27,28,29,30,31],[232,233],"Calypso Knowledge Base","Calypso Inbox","intermediate","Calypso","2026-05-23T11:03:21.530Z","/en/workflows/signal-reliability-navigator",{"en":237},{"title":9,"description":240,"ogDescription":241,"twitterDescription":241,"canonicalPath":237,"robots":242,"schemaType":243,"alternates":244},"Turn messy branch signals into trustworthy decisions with quick checks for dirty data, noisy metrics, and when to trust automation.","A practical decision workflow: spot polished noise, catch dirty signals early, compare branches safely, and know when automation needs human judgment.","index,follow","HowTo",[245],{"hreflang":6,"href":237},1780761213397]