[{"data":1,"prerenderedAt":277},["ShallowReactive",2],{"/en/workflows/branch-signal-reliability-advisor":3},{"id":4,"slug":5,"locale":6,"translationGroupId":7,"localeSwitchApproved":8,"title":9,"description":10,"documentationMarkdown":11,"workflowJson":12,"category":256,"tags":257,"integrations":261,"difficulty":264,"author":265,"verified":32,"featured":32,"date":266,"modified":266,"icon":7,"imageSrc":7,"path":267,"alternates":268,"seo":269},"d058c1e1-9bb7-4cd4-a152-5ae97f2bf199","branch-signal-reliability-advisor","en",null,true,"Branch Signal Reliability Advisor","An interactive decision coach that helps teams judge which branch metrics to trust, spot dirty signals early, and choose when to rely on automation vs. human judgment.","## How it works\nThis workflow turns “we have numbers” into “we have decisions we can defend.” It starts with a Knowledge Base pass for quick answers, then guides the user through a short menu of decision-shaped signal checks—focused on branch performance, conversations, and attribution.\n\nIt’s designed for the real world: where bad data often looks perfectly professional until someone presents it with confidence. The workflow helps operators spot polished noise, keep messy evidence honest, and escalate to a human when the situation needs judgment.\n\n## Key features\n- Knowledge Base-first responses for fast, consistent guidance before routing logic runs\n- A simple WhatsApp button menu that keeps the conversation decision-shaped (not academic)\n- Practical “trust tests” for branch numbers, attribution comparisons, and conversation-driven signals\n- A single “Next steps” prompt that loops back to the menu or escalates to a human\n- Built-in handoff to an Insights Desk when the signal is high-stakes, time-sensitive, or ambiguous\n\n## Step-by-step\n1. **Trigger:** The workflow starts when a message enters the flow.\n2. **Knowledge Base policy:** Calypso attempts to answer using your Knowledge Base. If it can’t, the flow continues to guided prompts.\n3. **Choose a coaching track:** The user picks one option from a WhatsApp button menu:\n   - Which branch numbers deserve trust\n   - How to spot dirty signal early\n   - When to trust automation vs. human judgment\n   - How to turn messy evidence into usable insight\n   - What teams misread in comparisons & attribution\n   - How to build a signal culture that leads to decisions\n4. **Get a decision-ready response:** The workflow sends a concise checklist and “what usually goes wrong first” guidance for the chosen track.\n5. **Next steps:** The user chooses:\n   - **Back to menu** (loop to the main menu), or\n   - **Talk to Insights Desk** (handoff to a human).\n\n## Setup requirements\n- **WhatsApp channel** connected to Calypso (for interactive button messages).\n- **Calypso Knowledge Base** enabled and populated with your internal definitions (metrics, branch taxonomy, attribution rules, data sources).\n- **Insights Desk handoff destination** (set the department name/ID in the fallback node).\n- No additional credentials are required beyond your Calypso channel and Knowledge Base access.",{"id":5,"teamId":13,"name":9,"version":14,"workflowVersion":15,"nodes":16,"connections":202,"routingEnabled":8,"active":32},"calypso-public-library","1.0.0",1,[17,33,40,52,82,91,97,103,109,115,121,128,134,140,146,152,158,173,178,185,195],{"id":18,"name":19,"type":20,"typeVersion":15,"position":21,"parameters":24,"category":31,"deletable":32,"connectable":32},"cfg1","Workflow settings","flow-configs",[22,23],80,40,{"name":9,"description":25,"tags":26,"triggerType":30},"Guided WhatsApp decision coach for evaluating branch signals, attribution comparisons, and when automation needs human judgment.",[27,28,29],"signal-quality","branch-metrics","decision-support","input","policy",false,{"id":34,"name":35,"type":30,"typeVersion":15,"position":36,"parameters":39,"category":30,"deletable":32,"connectable":8},"in1","Inbound message",[37,38],120,200,{},{"id":41,"name":42,"type":43,"typeVersion":15,"position":44,"parameters":46,"category":51,"deletable":8,"connectable":8},"kb1","Knowledge Base: signal & decision guidance","knowledge-base-policy",[45,38],340,{"enabled":8,"fallbackToRouting":8,"sticky":8,"stickyMode":47,"activationOpener":48,"personalization":50},"default",{"enabled":32,"instruction":49},"",{"useContactName":32},"response",{"id":53,"name":54,"type":55,"typeVersion":15,"position":56,"parameters":58,"category":51,"deletable":8,"connectable":8},"menu1","Choose what you’re deciding","interactive-message",[57,38],560,{"messageType":59,"headerText":60,"bodyText":61,"footerText":62,"sectionTitle":49,"buttons":63,"ctaDisplayText":49,"ctaUrl":49},"button","Signal checks (fast, practical)","Pick the decision you’re trying to make. I’ll give you the 3–5 checks that catch wrong-but-confident conclusions.","Tip: If it feels \"too clean,\" it probably is.",[64,67,70,73,76,79],{"id":65,"title":66},"TRUST_NUMBERS","Trust branch #s",{"id":68,"title":69},"DIRTY_SIGNAL","Dirty signal",{"id":71,"title":72},"AUTO_VS_HUMAN","Auto vs human",{"id":74,"title":75},"MESSY_EVIDENCE","Messy to insight",{"id":77,"title":78},"MISREAD_COMPARE","Attribution traps",{"id":80,"title":81},"SIGNAL_CULTURE","Signal culture",{"id":83,"name":84,"type":85,"typeVersion":15,"position":86,"parameters":88,"category":90,"deletable":8,"connectable":8},"if1","If: trust branch numbers","if",[87,38],780,{"buttonId":65,"operator":89},"equals","routing",{"id":92,"name":93,"type":85,"typeVersion":15,"position":94,"parameters":96,"category":90,"deletable":8,"connectable":8},"if2","If: spot dirty signal",[95,38],980,{"buttonId":68,"operator":89},{"id":98,"name":99,"type":85,"typeVersion":15,"position":100,"parameters":102,"category":90,"deletable":8,"connectable":8},"if3","If: automation vs judgment",[101,38],1180,{"buttonId":71,"operator":89},{"id":104,"name":105,"type":85,"typeVersion":15,"position":106,"parameters":108,"category":90,"deletable":8,"connectable":8},"if4","If: messy evidence to insight",[107,38],1380,{"buttonId":74,"operator":89},{"id":110,"name":111,"type":85,"typeVersion":15,"position":112,"parameters":114,"category":90,"deletable":8,"connectable":8},"if5","If: comparison & attribution traps",[113,38],1580,{"buttonId":77,"operator":89},{"id":116,"name":117,"type":85,"typeVersion":15,"position":118,"parameters":120,"category":90,"deletable":8,"connectable":8},"if6","If: build signal culture",[119,38],1780,{"buttonId":80,"operator":89},{"id":122,"name":123,"type":124,"typeVersion":15,"position":125,"parameters":126,"category":51,"deletable":8,"connectable":8},"t1","Guidance: which branch numbers to trust","text-message",[87,45],{"text":127},"Branch numbers that deserve trust usually have *friction*—they can be reconciled, explained, and repeated.\n\nTrust it more when:\n1) It ties to a concrete event (transaction, appointment kept, inventory movement)—not “engagement.”\n2) Definitions don’t change mid-quarter (what counts as a lead, a visit, a conversion).\n3) You can explain variance with a short causal story *and* a quick sanity check (staffing, hours, stock-outs, local promos).\n4) It matches at least one independent signal (e.g., revenue vs. unit volume; calls vs. appointments).\n\nPolished noise patterns:\n- A branch ‘improves’ right after a tracking change.\n- Every branch moves in the same direction at the same time.\n- The metric gets cleaner as the stakes get higher.\n\nIf you only have one check: ask, “What would have to be true for this number to be wrong?” Then go look for *that*.",{"id":129,"name":130,"type":124,"typeVersion":15,"position":131,"parameters":132,"category":51,"deletable":8,"connectable":8},"t2","Guidance: spot dirty signal early",[95,45],{"text":133},"Dirty signal rarely announces itself. It shows up as *confidence without cost*.\n\nEarly warning checks before the meeting goes off the rails:\n1) **Edge spikes:** A jump that starts exactly on a Monday / month boundary / system release date.\n2) **Too-perfect distributions:** Suddenly fewer outliers, fewer messy notes, fewer “unknowns.” Reality is untidy.\n3) **Missing denominator:** People quote totals without exposure (per staff hour, per footfall, per active customers).\n4) **Backfilled history:** Last month changes after this month’s tracking update.\n5) **Selective visibility:** One dashboard says ‘up’, the raw feed says ‘hmm.’\n\nPractical move: pick one branch, one day, and trace the metric to the source event. If you can’t do that in 10 minutes, you’re not holding a metric—you’re holding a story.",{"id":135,"name":136,"type":124,"typeVersion":15,"position":137,"parameters":138,"category":51,"deletable":8,"connectable":8},"t3","Guidance: automation vs judgment",[101,45],{"text":139},"Automation is great at *consistent rules*. Humans are better at *exception handling and ethics*.\n\nTrust automation when:\n- The decision is reversible (you can undo it cheaply).\n- The inputs are stable and well-defined (same source, same meaning).\n- The cost of delay is higher than the cost of being slightly wrong.\n\nRequire human judgment when:\n- The decision is irreversible or high-stakes (staffing cuts, compliance, major budget shifts).\n- The signal is gameable (teams can ‘optimize’ the metric instead of the outcome).\n- The data is proxy-heavy (sentiment, inferred intent, attribution models).\n\nA good rule: automate the *default*, and reserve people for the *deltas*—the weird cases, the conflicting signals, and the “this feels off” moments.",{"id":141,"name":142,"type":124,"typeVersion":15,"position":143,"parameters":144,"category":51,"deletable":8,"connectable":8},"t4","Guidance: messy evidence to usable insight",[107,45],{"text":145},"Cleaning can accidentally erase the truth—especially when the mess *is the signal*.\n\nHow to make messy evidence usable without sanding it down:\n1) **Keep the raw layer:** Always preserve verbatims, timestamps, and “unknown/other.” Don’t force-fit.\n2) **Tag, don’t rewrite:** Add lightweight labels (issue type, outcome, urgency) without changing original wording.\n3) **Separate ‘counting’ from ‘meaning’:** First quantify what happened; then interpret why.\n4) **Look for contradiction clusters:** The most valuable insight is often “the data disagrees,” because it points to process breaks.\n\nIf you must summarize: show ranges and caveats. A precise lie beats an honest range in a slide deck—until it doesn’t.",{"id":147,"name":148,"type":124,"typeVersion":15,"position":149,"parameters":150,"category":51,"deletable":8,"connectable":8},"t5","Guidance: comparison & attribution traps",[113,45],{"text":151},"Teams misread comparisons because they compare *outputs* while ignoring *inputs*.\n\nCommon traps (and the quick fix):\n1) **Different exposure:** Branch A has more footfall, better hours, different catchment. Fix: normalize (per staff hour, per visit, per qualified lead).\n2) **Mix shift:** Product/segment mix changes look like performance. Fix: compare like-for-like cohorts.\n3) **Attribution inflation:** When tracking improves, ‘marketing performance’ mysteriously improves too. Fix: annotate model/tracking changes; compare with holdouts or independent signals.\n4) **Conversation bias:** More conversations can mean more problems, not better service. Fix: separate inbound issue traffic from proactive outreach.\n\nIf you’re ranking branches, add one sentence: “What differs between branches that *isn’t* performance?” If you can’t answer, you’re ranking context.",{"id":153,"name":154,"type":124,"typeVersion":15,"position":155,"parameters":156,"category":51,"deletable":8,"connectable":8},"t6","Guidance: build a signal culture",[119,45],{"text":157},"A healthy signal culture doesn’t produce more dashboards. It produces faster, calmer decisions.\n\nMake it real:\n1) **Define ‘decision metrics’ vs ‘story metrics’:** Decision metrics change actions; story metrics fill slides.\n2) **Reward surfacing uncertainty:** If people get punished for caveats, they’ll bring you confident nonsense.\n3) **Track tracking changes:** Maintain a simple log of definition/model/source changes. Most “mystery trends” live there.\n4) **Use pre-mortems:** Before acting, ask: “How would we know this decision was wrong in 2 weeks?” Then instrument that.\n\nWit-sized truth: the first thing to break is usually not the model—it’s the incentives around the model.",{"id":159,"name":160,"type":55,"typeVersion":15,"position":161,"parameters":163,"category":51,"deletable":8,"connectable":8},"next1","Next step",[101,162],520,{"messageType":59,"headerText":160,"bodyText":164,"footerText":165,"sectionTitle":49,"buttons":166,"ctaDisplayText":49,"ctaUrl":49},"Want another signal check, or should we route this to a human for a quick read?","If the decision is high-stakes, escalate early.",[167,170],{"id":168,"title":169},"BACK_MENU","Another topic",{"id":171,"title":172},"TALK_TO_INSIGHTS","Insights Desk",{"id":174,"name":175,"type":85,"typeVersion":15,"position":176,"parameters":177,"category":90,"deletable":8,"connectable":8},"ifBack","If: another topic",[107,162],{"buttonId":168,"operator":89},{"id":179,"name":180,"type":124,"typeVersion":15,"position":181,"parameters":183,"category":51,"deletable":8,"connectable":8},"tRestart","How to get the menu again",[113,182],500,{"text":184},"To pick another topic, just send a new message (anything works). I’ll bring back the signal-check menu.\n\nIf this decision is time-sensitive or high-stakes, choose **Insights Desk** instead.",{"id":186,"name":187,"type":188,"typeVersion":15,"position":189,"parameters":191,"category":194,"deletable":8,"connectable":8},"handoff1","Handoff: Insights Desk","fallback",[113,190],620,{"handoffMessage":192,"departmentId":193,"departmentName":172},"Routing you to the Insights Desk. Please share: the branch(es), time window, the metric in question, and what decision you’re about to make. If there’s a tracking/model change, mention it—those are frequent ‘ghost trends.’","insights-desk","terminal",{"id":196,"name":197,"type":124,"typeVersion":15,"position":198,"parameters":200,"category":51,"deletable":8,"connectable":8},"tUnknown","If nothing matched",[199,38],1980,{"text":201},"I didn’t catch that selection. Please choose one of the buttons so I can give the right checks.",[203,208,210,212,215,218,220,222,224,226,228,230,232,234,236,238,240,242,244,246,248,250,252,254],{"id":204,"source":34,"target":41,"sourceHandle":205,"targetHandle":206,"type":207},"c1","out","in","smoothstep",{"id":209,"source":41,"target":53,"sourceHandle":205,"targetHandle":206,"type":207},"c2",{"id":211,"source":53,"target":83,"sourceHandle":205,"targetHandle":206,"type":207},"c3",{"id":213,"source":83,"target":122,"sourceHandle":214,"targetHandle":206,"type":207},"c4","true",{"id":216,"source":83,"target":92,"sourceHandle":217,"targetHandle":206,"type":207},"c5","false",{"id":219,"source":92,"target":129,"sourceHandle":214,"targetHandle":206,"type":207},"c6",{"id":221,"source":92,"target":98,"sourceHandle":217,"targetHandle":206,"type":207},"c7",{"id":223,"source":98,"target":135,"sourceHandle":214,"targetHandle":206,"type":207},"c8",{"id":225,"source":98,"target":104,"sourceHandle":217,"targetHandle":206,"type":207},"c9",{"id":227,"source":104,"target":141,"sourceHandle":214,"targetHandle":206,"type":207},"c10",{"id":229,"source":104,"target":110,"sourceHandle":217,"targetHandle":206,"type":207},"c11",{"id":231,"source":110,"target":147,"sourceHandle":214,"targetHandle":206,"type":207},"c12",{"id":233,"source":110,"target":116,"sourceHandle":217,"targetHandle":206,"type":207},"c13",{"id":235,"source":116,"target":153,"sourceHandle":214,"targetHandle":206,"type":207},"c14",{"id":237,"source":116,"target":196,"sourceHandle":217,"targetHandle":206,"type":207},"c15",{"id":239,"source":122,"target":159,"sourceHandle":205,"targetHandle":206,"type":207},"c16",{"id":241,"source":129,"target":159,"sourceHandle":205,"targetHandle":206,"type":207},"c17",{"id":243,"source":135,"target":159,"sourceHandle":205,"targetHandle":206,"type":207},"c18",{"id":245,"source":141,"target":159,"sourceHandle":205,"targetHandle":206,"type":207},"c19",{"id":247,"source":147,"target":159,"sourceHandle":205,"targetHandle":206,"type":207},"c20",{"id":249,"source":153,"target":159,"sourceHandle":205,"targetHandle":206,"type":207},"c21",{"id":251,"source":159,"target":174,"sourceHandle":205,"targetHandle":206,"type":207},"c23",{"id":253,"source":174,"target":179,"sourceHandle":214,"targetHandle":206,"type":207},"c24",{"id":255,"source":174,"target":186,"sourceHandle":217,"targetHandle":206,"type":207},"c25","automation",[27,28,29,258,259,260],"attribution","data-hygiene","ops-coaching",[262,263],"WhatsApp","Calypso Knowledge Base","intermediate","Calypso","2026-04-25T11:04:07.714Z","/en/workflows/branch-signal-reliability-advisor",{"en":267},{"title":9,"description":270,"ogDescription":271,"twitterDescription":272,"canonicalPath":267,"robots":273,"schemaType":274,"alternates":275},"Coach teams to trust the right branch signals, spot dirty data early, and decide when automation needs human judgment—via guided prompts.","A practical workflow to sanity check branch metrics, detect polished noise, and guide decisions with prompts—plus an Insights Desk handoff when needed.","Turn messy branch signals into decisions you can defend: trust tests, dirty signal checks, automation vs. judgment prompts, and a clean human handoff.","index,follow","HowTo",[276],{"hreflang":6,"href":267},1778614431975]