[{"data":1,"prerenderedAt":290},["ShallowReactive",2],{"/en/workflows/branch-data-trust-triage":3},{"id":4,"slug":5,"locale":6,"translationGroupId":7,"localeSwitchApproved":8,"title":9,"description":10,"documentationMarkdown":11,"workflowJson":12,"category":270,"tags":271,"integrations":274,"difficulty":277,"author":278,"verified":34,"featured":34,"date":279,"modified":279,"icon":7,"imageSrc":7,"path":280,"alternates":281,"seo":282},"7a28daf3-95e4-4c13-9459-aa271916c7c2","branch-data-trust-triage","en",null,true,"Branch Data Trust Triage","An advisor-style chat workflow that helps teams separate trustworthy branch signals from polished noise, spot dirty data early, and choose when to automate versus when to pause for human judgment.","## How it works\nThis workflow turns messy branch numbers, conversation notes, and “looks right” dashboards into decision-ready guidance—without pretending the data is cleaner than it is. It first tries to answer questions from your Calypso Knowledge Base, then routes people through a set of decision-shaped prompts that reveal where signal quality usually breaks.\n\nIt’s built for the moment right before a confident meeting goes off the rails: when everyone has charts, nobody has definitions, and the wrong metric is about to become a target. The workflow helps operators and leaders pressure-test signals quickly, choose the right level of automation, and escalate to a human analyst when the situation calls for judgment.\n\n## Key features\n- Knowledge Base-first answers for common signal and measurement questions, with smart routing when no article fits.\n- Button-driven triage menu focused on real decisions: trust, hygiene, attribution, and automation boundaries.\n- Practical “what to verify next” guidance that flags polished noise and common branch-comparison traps.\n- A second menu to either loop back for another check or hand off to an analyst team.\n- Built-in escalation path to a human when signals are ambiguous or stakes are high.\n\n## Step-by-step\n1. **Trigger:** A conversation starts (Input).\n2. **Knowledge Base assist:** The workflow attempts to answer using your **Calypso Knowledge Base**. If it can’t find a confident match, it continues to the menu.\n3. **Choose a check:** An interactive menu asks what kind of decision support the user needs (e.g., trusting branch numbers, spotting dirty signals, attribution comparisons).\n4. **Route by selection:** A matching **If** node detects the selected button and sends a tailored guidance message.\n5. **Next action menu:** A second interactive menu offers to **return to the main menu** or **talk to an analyst**.\n6. **Escalation (optional):** If the user requests help, the workflow hands off to the **Analytics Operations** queue with context.\n\n## Setup requirements\n- **Calypso Knowledge Base:** Recommended (not required). Add or curate articles on metric definitions, attribution rules, branch reporting cadence, and known data-quality pitfalls.\n- **Calypso Conversations/Inbox routing:** Ensure you have a department/queue for **Analytics Operations** (or equivalent) to receive handoffs.\n- **Credentials:** No external credentials are required for this workflow.",{"id":13,"teamId":14,"name":9,"version":15,"workflowVersion":16,"nodes":17,"connections":212,"routingEnabled":8,"active":34},"wf_branch_data_trust_triage","calypso-public-library","1.0.0",1,[18,35,41,53,84,93,101,107,113,119,125,131,137,143,149,155,161,176,182,189,195,205],{"id":19,"name":20,"type":21,"typeVersion":16,"position":22,"parameters":25,"category":33,"deletable":34,"connectable":34},"node_flow_configs","Workflow settings","flow-configs",[23,24],-80,-120,{"name":9,"description":26,"tags":27,"triggerType":32},"Advisor-style triage for branch signal trust, dirty data detection, attribution pitfalls, and automation boundaries.",[28,29,30,31],"signal-design","data-quality","branch-analytics","decision-making","input","policy",false,{"id":36,"name":37,"type":32,"typeVersion":16,"position":38,"parameters":40,"category":32,"deletable":34,"connectable":8},"node_input","Conversation starts",[23,39],40,{},{"id":42,"name":43,"type":44,"typeVersion":16,"position":45,"parameters":47,"category":52,"deletable":8,"connectable":8},"node_kb_policy","Knowledge Base assist","knowledge-base-policy",[46,39],160,{"enabled":8,"fallbackToRouting":8,"sticky":8,"stickyMode":48,"activationOpener":49,"personalization":51},"default",{"enabled":8,"instruction":50},"Ask your question, or pick a quick check below to pressure-test what you’re seeing.",{"useContactName":8},"response",{"id":54,"name":55,"type":56,"typeVersion":16,"position":57,"parameters":59,"category":52,"deletable":8,"connectable":8},"node_menu_main","Pick a signal check","interactive-message",[58,39],420,{"messageType":60,"headerText":61,"bodyText":62,"footerText":63,"sectionTitle":64,"buttons":65,"ctaDisplayText":64,"ctaUrl":64},"button","Signal trust checks","What are you trying to decide right now? Pick the closest option and I’ll give you fast reality checks (the kind that prevent confident wrong decisions).","Tip: start with Trust branch nums.","",[66,69,72,75,78,81],{"id":67,"title":68},"trust_branch_numbers","Trust branch nums",{"id":70,"title":71},"spot_dirty_signal","Spot dirty signal",{"id":73,"title":74},"automation_vs_judgment","Auto vs judgment",{"id":76,"title":77},"messy_evidence","Use messy evidence",{"id":79,"title":80},"compare_branches_attribution","Branch & attribution",{"id":82,"title":83},"build_signal_culture","Signal culture",{"id":85,"name":86,"type":87,"typeVersion":16,"position":88,"parameters":90,"category":92,"deletable":8,"connectable":8},"node_if_trust","If: trust branch numbers","if",[89,24],680,{"buttonId":67,"operator":91},"equals","routing",{"id":94,"name":95,"type":96,"typeVersion":16,"position":97,"parameters":99,"category":52,"deletable":8,"connectable":8},"node_text_trust","Guidance: trust branch numbers","text-message",[98,24],940,{"text":100},"Branch numbers usually deserve trust when they’re *boring in the right ways*: consistent definitions, stable collection, and explainable drift.\n\nFast checks that separate signal from polished noise:\n- **Definition lock:** Can two people independently say what the metric includes/excludes and get the same answer?\n- **Denominator sanity:** Did the base change (active customers, eligible leads, staffed days)? Most “growth” bugs live here.\n- **Timing:** Are you looking at event time, posting time, or report refresh time? Pick the wrong one and you’ll ‘discover’ fake trends.\n- **Operational plausibility:** Does the magnitude fit the real-world constraint (capacity, hours, inventory)? If not, it’s measurement—not performance.\n\nWhen a number is pretty but untrustworthy, it usually has one of three smells: a silent definition change, missing records that look like zeros, or a metric that’s actually a proxy for process compliance rather than outcomes.",{"id":102,"name":103,"type":87,"typeVersion":16,"position":104,"parameters":106,"category":92,"deletable":8,"connectable":8},"node_if_dirty","If: spot dirty signal",[89,105],-20,{"buttonId":70,"operator":91},{"id":108,"name":109,"type":96,"typeVersion":16,"position":110,"parameters":111,"category":52,"deletable":8,"connectable":8},"node_text_dirty","Guidance: spot dirty signal",[98,105],{"text":112},"Dirty signal rarely announces itself. It shows up as *clean-looking certainty*.\n\nUse this pre-meeting checklist:\n- **Sharp discontinuities:** A step-change on a date that doesn’t match a real operational change is usually instrumentation.\n- **Too-perfect distributions:** If every branch suddenly clusters tightly, someone changed rounding, thresholds, or backfilled.\n- **Missingness that behaves:** Missing data that’s patterned (only weekends, only one branch, only one channel) will bias comparisons.\n- **Single-source truth:** If one system ‘wins’ every disagreement, you may have a reconciliation problem, not a performance story.\n\nPractical move: ask “What would have to be true operationally for this to be real?” If the answer is ‘a miracle,’ investigate collection and definitions before you debate strategy.",{"id":114,"name":115,"type":87,"typeVersion":16,"position":116,"parameters":118,"category":92,"deletable":8,"connectable":8},"node_if_auto","If: automation vs judgment",[89,117],80,{"buttonId":73,"operator":91},{"id":120,"name":121,"type":96,"typeVersion":16,"position":122,"parameters":123,"category":52,"deletable":8,"connectable":8},"node_text_auto","Guidance: automation vs judgment",[98,117],{"text":124},"Trust automation when the world is stable and the rules are explicit. Pull in humans when exceptions are common or incentives are sharp.\n\nGood candidates for automation:\n- **High-volume, low-stakes decisions** (routing, reminders, threshold alerts).\n- **Well-defined inputs** with consistent capture (timestamps, structured statuses).\n- **Clear failure modes** (you can quantify false positives/negatives).\n\nRed flags that require human judgment:\n- **Ambiguous labels** (‘qualified’, ‘interested’, ‘resolved’) that drift by branch or manager.\n- **Adversarial incentives** (metrics tied to bonus/targets). The data will learn to lie politely.\n- **Non-stationarity:** seasonality, new campaigns, staffing changes, process rollouts.\n\nRule of thumb: automate the *first pass* and put humans on the boundary cases—where the decision cost of being wrong is highest.",{"id":126,"name":127,"type":87,"typeVersion":16,"position":128,"parameters":130,"category":92,"deletable":8,"connectable":8},"node_if_messy","If: messy evidence",[89,129],180,{"buttonId":76,"operator":91},{"id":132,"name":133,"type":96,"typeVersion":16,"position":134,"parameters":135,"category":52,"deletable":8,"connectable":8},"node_text_messy","Guidance: messy evidence",[98,129],{"text":136},"Cleaning is useful—until it starts cleaning away the truth.\n\nTurn messy evidence into usable insight without sanding off reality:\n- **Preserve raw + derived:** Keep the messy original, then create a decision-ready view. Don’t overwrite.\n- **Track reasons, not just outcomes:** A ‘no-show’ vs ‘rescheduled’ matters; collapsing them makes dashboards pretty and decisions wrong.\n- **Prefer cohorts over averages:** Averages hide operational failures (new staff vs experienced, weekday vs weekend).\n- **Explain variance before optimizing:** If you can’t explain why branches differ, you’re not ready to copy the ‘best’ branch.\n\nPractical move: write a one-sentence decision question first (“Should we change staffing model X?”). Then keep only the evidence that changes that decision, and label what’s uncertain.",{"id":138,"name":139,"type":87,"typeVersion":16,"position":140,"parameters":142,"category":92,"deletable":8,"connectable":8},"node_if_compare","If: compare branches & attribution",[89,141],280,{"buttonId":79,"operator":91},{"id":144,"name":145,"type":96,"typeVersion":16,"position":146,"parameters":147,"category":52,"deletable":8,"connectable":8},"node_text_compare","Guidance: comparing branches & attribution",[98,141],{"text":148},"Branch comparisons go wrong in repeatable ways—mostly because branches aren’t actually comparable.\n\nWhat teams misread first:\n- **Mix shift:** Branch A ‘wins’ because it gets easier customers/leads, different hours, or different product mix.\n- **Capacity masking:** A high conversion rate can mean low volume (only working the best leads).\n- **Attribution double-counting:** The same customer touches multiple channels; you’ll ‘create’ ROI by counting one outcome multiple times.\n- **Process vs performance:** Branch with stricter logging looks worse because it records failure honestly.\n\nPractical move: before ranking branches, standardize on (1) eligible population, (2) time window, and (3) what counts as an attempt. If you can’t align those, you don’t have a leaderboard—you have a storytelling contest.",{"id":150,"name":151,"type":87,"typeVersion":16,"position":152,"parameters":154,"category":92,"deletable":8,"connectable":8},"node_if_culture","If: build signal culture",[89,153],380,{"buttonId":82,"operator":91},{"id":156,"name":157,"type":96,"typeVersion":16,"position":158,"parameters":159,"category":52,"deletable":8,"connectable":8},"node_text_culture","Guidance: build signal culture",[98,153],{"text":160},"A strong signal culture doesn’t produce more slides. It produces faster, calmer decisions.\n\nMake it real (and durable):\n- **One metric, one owner, one definition.** If ownership is shared, accountability is imaginary.\n- **Decision logs:** Record what you decided, why, and what you expected to happen. Reality will train your instincts.\n- **Red-team the dashboard:** Assign someone to argue the data is wrong *before* you argue what to do.\n- **Reward truth-telling:** Branches should not be punished for reporting problems—only for hiding them.\n\nWitty but true: the first casualty of a target is usually the measurement. Design incentives like you’ve met humans before.",{"id":162,"name":163,"type":56,"typeVersion":16,"position":164,"parameters":166,"category":52,"deletable":8,"connectable":8},"node_menu_next","Next step",[165,117],1200,{"messageType":60,"headerText":167,"bodyText":168,"footerText":64,"sectionTitle":64,"buttons":169,"ctaDisplayText":64,"ctaUrl":64},"What next?","Want another check, or do you need an analyst to look at a specific case (definitions, missingness, attribution, or a suspicious jump)?",[170,173],{"id":171,"title":172},"back_to_menu","Another signal check",{"id":174,"title":175},"talk_to_analyst","Talk to an analyst",{"id":177,"name":178,"type":87,"typeVersion":16,"position":179,"parameters":181,"category":92,"deletable":8,"connectable":8},"node_if_back","If: another signal check",[180,39],1460,{"buttonId":171,"operator":91},{"id":183,"name":184,"type":96,"typeVersion":16,"position":185,"parameters":187,"category":52,"deletable":8,"connectable":8},"node_text_restart","How to run another check",[186,39],1720,{"text":188},"To run another signal check, just ask a new question in this chat (or start a new conversation) and I’ll guide you again. If you’re heading into a high-stakes meeting, consider the analyst handoff instead—faster than debating a metric’s definition in real time.",{"id":190,"name":191,"type":87,"typeVersion":16,"position":192,"parameters":194,"category":92,"deletable":8,"connectable":8},"node_if_handoff","If: talk to analyst",[180,193],140,{"buttonId":174,"operator":91},{"id":196,"name":197,"type":198,"typeVersion":16,"position":199,"parameters":200,"category":204,"deletable":8,"connectable":8},"node_fallback_handoff","Handoff to Analytics Operations","fallback",[186,193],{"handoffMessage":201,"departmentId":202,"departmentName":203},"Got it. I’m handing this to Analytics Operations. To speed things up, please share: (1) the metric name + definition you’re using, (2) branches/time window, and (3) what looks ‘off’ (a jump, mismatch, or ranking).","analytics-ops","Analytics Operations","terminal",{"id":206,"name":207,"type":96,"typeVersion":16,"position":208,"parameters":210,"category":52,"deletable":8,"connectable":8},"node_text_next_invalid","Next step not selected",[186,209],240,{"text":211},"Please tap one of the buttons above so I can route you correctly (another check or analyst handoff).",[213,218,220,222,225,228,230,232,234,236,238,240,242,244,246,248,250,252,254,256,258,260,262,264,266,268],{"id":214,"source":36,"target":42,"sourceHandle":215,"targetHandle":216,"type":217},"conn_input_to_kb","out","in","edge",{"id":219,"source":42,"target":54,"sourceHandle":215,"targetHandle":216,"type":217},"conn_kb_to_menu",{"id":221,"source":54,"target":85,"sourceHandle":215,"targetHandle":216,"type":217},"conn_menu_to_if_trust",{"id":223,"source":85,"target":94,"sourceHandle":224,"targetHandle":216,"type":217},"conn_if_trust_true_to_text","true",{"id":226,"source":85,"target":102,"sourceHandle":227,"targetHandle":216,"type":217},"conn_if_trust_false_to_if_dirty","false",{"id":229,"source":102,"target":108,"sourceHandle":224,"targetHandle":216,"type":217},"conn_if_dirty_true_to_text",{"id":231,"source":102,"target":114,"sourceHandle":227,"targetHandle":216,"type":217},"conn_if_dirty_false_to_if_auto",{"id":233,"source":114,"target":120,"sourceHandle":224,"targetHandle":216,"type":217},"conn_if_auto_true_to_text",{"id":235,"source":114,"target":126,"sourceHandle":227,"targetHandle":216,"type":217},"conn_if_auto_false_to_if_messy",{"id":237,"source":126,"target":132,"sourceHandle":224,"targetHandle":216,"type":217},"conn_if_messy_true_to_text",{"id":239,"source":126,"target":138,"sourceHandle":227,"targetHandle":216,"type":217},"conn_if_messy_false_to_if_compare",{"id":241,"source":138,"target":144,"sourceHandle":224,"targetHandle":216,"type":217},"conn_if_compare_true_to_text",{"id":243,"source":138,"target":150,"sourceHandle":227,"targetHandle":216,"type":217},"conn_if_compare_false_to_if_culture",{"id":245,"source":150,"target":156,"sourceHandle":224,"targetHandle":216,"type":217},"conn_if_culture_true_to_text",{"id":247,"source":94,"target":162,"sourceHandle":215,"targetHandle":216,"type":217},"conn_text_trust_to_next",{"id":249,"source":108,"target":162,"sourceHandle":215,"targetHandle":216,"type":217},"conn_text_dirty_to_next",{"id":251,"source":120,"target":162,"sourceHandle":215,"targetHandle":216,"type":217},"conn_text_auto_to_next",{"id":253,"source":132,"target":162,"sourceHandle":215,"targetHandle":216,"type":217},"conn_text_messy_to_next",{"id":255,"source":144,"target":162,"sourceHandle":215,"targetHandle":216,"type":217},"conn_text_compare_to_next",{"id":257,"source":156,"target":162,"sourceHandle":215,"targetHandle":216,"type":217},"conn_text_culture_to_next",{"id":259,"source":150,"target":162,"sourceHandle":227,"targetHandle":216,"type":217},"conn_if_culture_false_to_next",{"id":261,"source":162,"target":177,"sourceHandle":215,"targetHandle":216,"type":217},"conn_next_to_if_back",{"id":263,"source":177,"target":183,"sourceHandle":224,"targetHandle":216,"type":217},"conn_if_back_true_to_restart_text",{"id":265,"source":177,"target":190,"sourceHandle":227,"targetHandle":216,"type":217},"conn_if_back_false_to_if_handoff",{"id":267,"source":190,"target":196,"sourceHandle":224,"targetHandle":216,"type":217},"conn_if_handoff_true_to_fallback",{"id":269,"source":190,"target":206,"sourceHandle":227,"targetHandle":216,"type":217},"conn_if_handoff_false_to_invalid_next","automation",[28,29,30,31,272,273],"attribution","triage",[275,276],"Calypso Knowledge Base","Calypso Conversations","intermediate","Calypso","2026-06-03T11:04:25.775Z","/en/workflows/branch-data-trust-triage",{"en":280},{"title":9,"description":283,"ogDescription":284,"twitterDescription":285,"canonicalPath":280,"robots":286,"schemaType":287,"alternates":288},"Route teams from messy branch signals to reliable decisions with KB answers, dirty signal checks, and analyst escalation when needed.","A practical workflow for branch signal trust: KB first answers, dirty data detection prompts, and a clean handoff to analysts when judgment matters.","Turn branch numbers and messy evidence into decision ready guidance: KB first responses, signal hygiene checks, and analyst escalation when needed.","index,follow","HowTo",[289],{"hreflang":6,"href":280},1780761212900]