[{"data":1,"prerenderedAt":281},["ShallowReactive",2],{"/en/workflows/trustworthy-signal-triage-for-leaders":3},{"id":4,"slug":5,"locale":6,"translationGroupId":7,"localeSwitchApproved":8,"title":9,"description":10,"documentationMarkdown":11,"workflowJson":12,"category":261,"tags":262,"integrations":266,"difficulty":268,"author":269,"verified":32,"featured":32,"date":270,"modified":270,"icon":7,"imageSrc":7,"path":271,"alternates":272,"seo":273},"ca07aab3-35d1-453f-98c7-c494f68610c3","trustworthy-signal-triage-for-leaders","en",null,true,"Trustworthy Signal Triage for Leaders","A guided decision-support flow that helps leaders and operators stress-test branch numbers, spot dirty signals early, and decide when automation is safe—or when human judgment must take over.","## How it works\nThis workflow turns “confident-sounding” branch numbers, conversations, and event signals into decision-ready judgment. It starts with your Knowledge Base for fast, consistent answers—then routes the user through a practical menu of reality checks that catch the common ways teams talk themselves into the wrong conclusion.\n\nIt’s built for messy, real-world evidence: partial attribution, inconsistent logging, and dashboards that look clean while the underlying signal is not. The tone is advisory (not academic), with clear prompts that help the team decide what to trust, what to measure next, and when to escalate.\n\n## Key features\n- Knowledge Base–first responses for consistent guardrails and shared definitions\n- Button-driven triage menu covering the highest-risk decision moments (branch comparisons, attribution, automation vs. judgment)\n- “Dirty signal” spotting prompts designed to prevent confident meetings from going off the rails\n- A single follow-up menu to keep the conversation moving (back to menu or escalate)\n- Optional human handoff path to an Analytics Support department\n\n## Step-by-step\n1. **Trigger:** A user starts the workflow (Input).\n2. **Knowledge Base pass:** The workflow checks your **Knowledge Base Policy** for relevant guidance first.\n3. **Decision menu:** The user chooses what they need help with (Interactive Message):\n   - Which branch numbers deserve trust\n   - Spot dirty signal before a meeting\n   - Automation vs human judgment\n   - Comparing branches & attribution\n   - Building a signal culture\n   - Talk to a human\n4. **Routing:** The workflow matches the selected button and sends the corresponding guidance (IF → Text Message).\n5. **Keep moving:** A follow-up menu lets the user **return to the main menu** or **handoff to a human** (Interactive Message → IF).\n6. **Escalation (optional):** If requested, the workflow hands off to **Analytics Support** (Fallback).\n\n## Setup requirements\n- **Calypso Knowledge Base**: Recommended. Add short, practical articles that match the menu items (branch metric definitions, logging standards, attribution pitfalls, automation guardrails).\n- **Department for handoff (optional):** Create or select an **Analytics Support** (or equivalent) department to receive escalations.\n- **Credentials:** None required for this workflow.",{"id":13,"teamId":14,"name":9,"version":15,"workflowVersion":16,"nodes":17,"connections":206,"routingEnabled":8,"active":32},"wf_signal_triage_leaders_v1","calypso-public-library","1.0.0",1,[18,33,39,51,82,92,98,104,110,116,122,130,136,142,148,154,169,176,182,189,196],{"id":19,"name":20,"type":21,"typeVersion":16,"position":22,"parameters":24,"category":31,"deletable":32,"connectable":32},"cfg_1","Workflow settings","flow-configs",[23,23],80,{"name":9,"description":25,"tags":26,"triggerType":30},"Knowledge-base-first triage for branch signal reliability, dirty signal detection, automation guardrails, and branch comparison pitfalls.",[27,28,29],"signal-triage","decision-quality","branch-metrics","input","policy",false,{"id":34,"name":35,"type":30,"typeVersion":16,"position":36,"parameters":38,"category":30,"deletable":32,"connectable":8},"in_1","Start",[23,37],220,{},{"id":40,"name":41,"type":42,"typeVersion":16,"position":43,"parameters":45,"category":50,"deletable":8,"connectable":8},"kb_1","Knowledge Base guardrails","knowledge-base-policy",[44,37],320,{"enabled":8,"fallbackToRouting":8,"sticky":8,"stickyMode":46,"activationOpener":47,"personalization":49},"default",{"enabled":8,"instruction":48},"Answer using the Knowledge Base when possible. Be practical and plainspoken: focus on what to trust, what to measure next, and what usually breaks first. Avoid academic jargon.",{"useContactName":8},"response",{"id":52,"name":53,"type":54,"typeVersion":16,"position":55,"parameters":57,"category":50,"deletable":8,"connectable":8},"menu_main_1","Signal triage menu","interactive-message",[56,37],580,{"messageType":58,"headerText":59,"bodyText":60,"footerText":61,"sectionTitle":62,"buttons":63,"ctaDisplayText":62,"ctaUrl":62},"button","Signal-to-decision triage","Pick what you’re deciding. I’ll help you separate solid signal from polished noise—before the meeting gets confident for the wrong reasons.","Choose one to continue.","",[64,67,70,73,76,79],{"id":65,"title":66},"trust_branch_numbers","Trust branch nums",{"id":68,"title":69},"spot_dirty_signal","Dirty signal check",{"id":71,"title":72},"automation_vs_judgment","Automation vs human",{"id":74,"title":75},"compare_branches_attribution","Branch compare",{"id":77,"title":78},"build_signal_culture","Signal culture",{"id":80,"title":81},"talk_to_human","Talk to human",{"id":83,"name":84,"type":85,"typeVersion":16,"position":86,"parameters":89,"category":91,"deletable":8,"connectable":8},"if_trust_1","If: trust branch numbers","if",[87,88],820,120,{"buttonId":65,"operator":90},"equals","routing",{"id":93,"name":94,"type":85,"typeVersion":16,"position":95,"parameters":97,"category":91,"deletable":8,"connectable":8},"if_dirty_1","If: spot dirty signal",[87,96],200,{"buttonId":68,"operator":90},{"id":99,"name":100,"type":85,"typeVersion":16,"position":101,"parameters":103,"category":91,"deletable":8,"connectable":8},"if_auto_1","If: automation vs judgment",[87,102],280,{"buttonId":71,"operator":90},{"id":105,"name":106,"type":85,"typeVersion":16,"position":107,"parameters":109,"category":91,"deletable":8,"connectable":8},"if_compare_1","If: compare branches & attribution",[87,108],360,{"buttonId":74,"operator":90},{"id":111,"name":112,"type":85,"typeVersion":16,"position":113,"parameters":115,"category":91,"deletable":8,"connectable":8},"if_culture_1","If: build signal culture",[87,114],440,{"buttonId":77,"operator":90},{"id":117,"name":118,"type":85,"typeVersion":16,"position":119,"parameters":121,"category":91,"deletable":8,"connectable":8},"if_human_main_1","If: talk to a human (main)",[87,120],520,{"buttonId":80,"operator":90},{"id":123,"name":124,"type":125,"typeVersion":16,"position":126,"parameters":128,"category":50,"deletable":8,"connectable":8},"txt_trust_1","Guidance: which numbers to trust","text-message",[127,88],1080,{"text":129},"How to tell if a branch number deserves trust (or is just well-presented noise):\n\n1) Start with **definition stability**: does everyone mean the same thing by the metric across branches and weeks? If the definition “drifts,” the trend is fiction.\n2) Check **capture mechanics**: is the number produced by observed events (good) or by manually keyed fields and optional tags (usually fragile)?\n3) Look for **incentive pressure**: anything tied to targets attracts “creative compliance.” The cleaner the story, the more you should ask what was excluded.\n4) Demand **denominator clarity**: counts without exposure/context (traffic, eligible customers, baseline) are confidence theater.\n5) Validate **edge cases**: refunds, cancellations, reopened cases, duplicates—these are where branches silently diverge.\n\nPractical rule: if a metric can’t survive a 5-minute audit (“show me three raw examples end-to-end”), don’t let it drive a confident decision.",{"id":131,"name":132,"type":125,"typeVersion":16,"position":133,"parameters":134,"category":50,"deletable":8,"connectable":8},"txt_dirty_1","Guidance: spot dirty signal",[127,96],{"text":135},"Dirty signal usually looks fine—right up until it wins an argument in a meeting.\n\nQuick ways to spot it before it’s expensive:\n- **Sudden step-changes** that line up with process changes, staffing changes, or a new form field (not customer behavior).\n- **Too-perfect symmetry** across branches (often templated reporting, not reality).\n- **Missingness that isn’t random**: the hardest cases are the first to stop being logged.\n- **Conversation summaries that always agree** with the outcome (that’s not insight; that’s hindsight).\n- **Attribution that can’t name its losers**: if every channel “drives impact,” you’re measuring coverage, not causality.\n\nMeeting-saver question: “What would we expect to see if this were an instrumentation artifact?” If nobody can answer, you’re not ready to decide.",{"id":137,"name":138,"type":125,"typeVersion":16,"position":139,"parameters":140,"category":50,"deletable":8,"connectable":8},"txt_auto_1","Guidance: automation vs judgment",[127,102],{"text":141},"When to trust automation vs when you still need human judgment:\n\nTrust automation when:\n- The input signal is **consistent, observable, and audited** (clicks, timestamps, validated events).\n- The downside of a wrong decision is **small and reversible**.\n- You have **drift checks** (you’ll notice when behavior changes).\n\nInsist on human judgment when:\n- The metric is **proxy-heavy** (sentiment scores, “intent,” self-reported reasons).\n- Branches differ in **how work is done**, not just volume.\n- The decision is **one-way** (staffing cuts, policy changes, compliance actions).\n\nPractical compromise: automate the default, but add a human gate for outliers, low-confidence cases, and anything that affects customers asymmetrically.",{"id":143,"name":144,"type":125,"typeVersion":16,"position":145,"parameters":146,"category":50,"deletable":8,"connectable":8},"txt_compare_1","Guidance: comparing branches & attribution",[127,108],{"text":147},"What teams misread first when comparing branches (and why attribution lies politely):\n\n1) **Different denominators**: Branch A’s ‘conversion’ and Branch B’s might be built on different eligibility pools.\n2) **Case-mix**: one branch gets the easy work; the other gets the weird stuff. Your metric is a workload proxy unless you adjust.\n3) **Timing artifacts**: backlog, weekends, staffing patterns—these create “performance” differences with no underlying quality change.\n4) **Conversation volume ≠ conversation quality**: more chats can mean more confusion, not better engagement.\n5) **Attribution inflation**: if multiple touches get credit and no one owns the null, every channel looks heroic.\n\nDecision-safe approach: compare branches only after you can explain (a) who was eligible, (b) what was logged, and (c) what changed operationally.",{"id":149,"name":150,"type":125,"typeVersion":16,"position":151,"parameters":152,"category":50,"deletable":8,"connectable":8},"txt_culture_1","Guidance: build signal culture",[127,114],{"text":153},"How to build a signal culture that helps decisions happen (not just slides):\n\n- Create a short list of **decision-grade metrics** with explicit owners, definitions, and known failure modes.\n- Make the first question in reviews: **“What would change our mind?”** If nothing can, you’re doing theater.\n- Reward teams for **finding instrumentation bugs**, not for defending dashboards.\n- Keep a “**metric obituary**” log: when a metric misleads, write down how, and what guardrail you added.\n- Default to **small, reversible tests** when evidence is messy.\n\nWit-with-teeth rule: if a metric only looks good in a deck and not in the raw examples, it’s a design problem—not a performance problem.",{"id":155,"name":156,"type":54,"typeVersion":16,"position":157,"parameters":160,"category":50,"deletable":8,"connectable":8},"menu_follow_1","Next step",[158,159],1340,300,{"messageType":58,"headerText":161,"bodyText":162,"footerText":163,"sectionTitle":62,"buttons":164,"ctaDisplayText":62,"ctaUrl":62},"Keep going","Want to check another angle, or escalate to a human?","Choose one.",[165,168],{"id":166,"title":167},"another_topic","Another topic",{"id":80,"title":81},{"id":170,"name":171,"type":85,"typeVersion":16,"position":172,"parameters":175,"category":91,"deletable":8,"connectable":8},"if_another_1","If: another topic",[173,174],1600,260,{"buttonId":166,"operator":90},{"id":177,"name":178,"type":85,"typeVersion":16,"position":179,"parameters":181,"category":91,"deletable":8,"connectable":8},"if_human_follow_1","If: talk to a human (follow-up)",[173,180],340,{"buttonId":80,"operator":90},{"id":183,"name":184,"type":125,"typeVersion":16,"position":185,"parameters":187,"category":50,"deletable":8,"connectable":8},"txt_restart_1","Restart note",[186,174],1860,{"text":188},"To explore another topic, restart this workflow and pick a different button from the menu. (Yes, it’s one extra click—cheaper than one confident wrong decision.)",{"id":190,"name":191,"type":125,"typeVersion":16,"position":192,"parameters":194,"category":50,"deletable":8,"connectable":8},"txt_unknown_1","Unknown selection",[127,193],600,{"text":195},"I didn’t catch that selection. Please restart the workflow and choose one of the buttons so I can route you to the right reality check.",{"id":197,"name":198,"type":199,"typeVersion":16,"position":200,"parameters":201,"category":205,"deletable":8,"connectable":8},"handoff_1","Handoff to Analytics Support","fallback",[158,120],{"handoffMessage":202,"departmentId":203,"departmentName":204},"Got it. I’m handing this to Analytics Support so you can sanity-check the signal with a human. If you can, share: the metric definition, timeframe, and 2–3 raw examples end-to-end.","dept-analytics-support","Analytics Support","terminal",[207,211,213,215,218,221,223,225,227,229,231,233,235,237,239,241,243,245,247,249,251,253,255,257,259],{"id":208,"source":34,"target":40,"sourceHandle":209,"targetHandle":210,"type":46},"c_in_kb","out","in",{"id":212,"source":40,"target":52,"sourceHandle":209,"targetHandle":210,"type":46},"c_kb_menu",{"id":214,"source":52,"target":83,"sourceHandle":209,"targetHandle":210,"type":46},"c_menu_if_trust",{"id":216,"source":83,"target":123,"sourceHandle":217,"targetHandle":210,"type":46},"c_if_trust_true","true",{"id":219,"source":83,"target":93,"sourceHandle":220,"targetHandle":210,"type":46},"c_if_trust_false","false",{"id":222,"source":93,"target":131,"sourceHandle":217,"targetHandle":210,"type":46},"c_if_dirty_true",{"id":224,"source":93,"target":99,"sourceHandle":220,"targetHandle":210,"type":46},"c_if_dirty_false",{"id":226,"source":99,"target":137,"sourceHandle":217,"targetHandle":210,"type":46},"c_if_auto_true",{"id":228,"source":99,"target":105,"sourceHandle":220,"targetHandle":210,"type":46},"c_if_auto_false",{"id":230,"source":105,"target":143,"sourceHandle":217,"targetHandle":210,"type":46},"c_if_compare_true",{"id":232,"source":105,"target":111,"sourceHandle":220,"targetHandle":210,"type":46},"c_if_compare_false",{"id":234,"source":111,"target":149,"sourceHandle":217,"targetHandle":210,"type":46},"c_if_culture_true",{"id":236,"source":111,"target":117,"sourceHandle":220,"targetHandle":210,"type":46},"c_if_culture_false",{"id":238,"source":117,"target":197,"sourceHandle":217,"targetHandle":210,"type":46},"c_if_human_main_true",{"id":240,"source":117,"target":190,"sourceHandle":220,"targetHandle":210,"type":46},"c_if_human_main_false",{"id":242,"source":123,"target":155,"sourceHandle":209,"targetHandle":210,"type":46},"c_txt_trust_follow",{"id":244,"source":131,"target":155,"sourceHandle":209,"targetHandle":210,"type":46},"c_txt_dirty_follow",{"id":246,"source":137,"target":155,"sourceHandle":209,"targetHandle":210,"type":46},"c_txt_auto_follow",{"id":248,"source":143,"target":155,"sourceHandle":209,"targetHandle":210,"type":46},"c_txt_compare_follow",{"id":250,"source":149,"target":155,"sourceHandle":209,"targetHandle":210,"type":46},"c_txt_culture_follow",{"id":252,"source":155,"target":170,"sourceHandle":209,"targetHandle":210,"type":46},"c_follow_if_another",{"id":254,"source":170,"target":183,"sourceHandle":217,"targetHandle":210,"type":46},"c_if_another_true",{"id":256,"source":170,"target":177,"sourceHandle":220,"targetHandle":210,"type":46},"c_if_another_false",{"id":258,"source":177,"target":197,"sourceHandle":217,"targetHandle":210,"type":46},"c_if_human_follow_true",{"id":260,"source":177,"target":183,"sourceHandle":220,"targetHandle":210,"type":46},"c_if_human_follow_false","automation",[27,29,28,263,264,265],"data-hygiene","automation-guardrails","attribution",[267],"Calypso Knowledge Base","intermediate","Calypso","2026-06-19T11:04:04.694Z","/en/workflows/trustworthy-signal-triage-for-leaders",{"en":271},{"title":9,"description":274,"ogDescription":275,"twitterDescription":276,"canonicalPath":271,"robots":277,"schemaType":278,"alternates":279},"Guide leaders to trust the right branch signals, spot dirty data early, and choose automation vs human judgment with a structured triage menu.","A practical triage flow that stress tests branch numbers, catches dirty signal before decisions, and clarifies when automation is safe vs when humans must step in.","Turn messy branch signals into decision ready judgment: spot dirty data, compare branches safely, and know when automation needs a human override.","index,follow","HowTo",[280],{"hreflang":6,"href":271},1785947669077]