[{"data":1,"prerenderedAt":261},["ShallowReactive",2],{"/en/workflows/signal-reliability-navigator-for-branch-decisions":3},{"id":4,"slug":5,"locale":6,"translationGroupId":7,"localeSwitchApproved":8,"title":9,"description":10,"documentationMarkdown":11,"workflowJson":12,"category":241,"tags":242,"integrations":246,"difficulty":248,"author":249,"verified":33,"featured":33,"date":250,"modified":250,"icon":7,"imageSrc":7,"path":251,"alternates":252,"seo":253},"079b5a5c-6485-4314-ae66-47310af9a67f","signal-reliability-navigator-for-branch-decisions","en",null,true,"Signal Reliability Navigator for Branch Decisions","A guided chat workflow that helps teams separate trustworthy branch signals from polished noise, spot dirty data early, and decide when automation needs human judgment.","## How it works\nThis workflow turns “we have numbers” into “we have decisions.” When someone pings your Calypso Inbox with a messy branch question (metrics, conversations, attribution, operational events), the workflow first tries to answer from your Knowledge Base. If it can’t confidently resolve the question, it routes the person into a set of decision-shaped mini playbooks.\n\nEach path is built to catch the failure mode that usually bites first: clean-looking data that’s quietly inconsistent, comparability traps across branches, and automation that’s technically correct while practically misleading. It’s designed to keep confident meetings from going confidently wrong.\n\n## Key features\n- Knowledge Base first, routing second: answers known questions fast, escalates the rest.\n- Button-based menu of practical decision checks (trustworthy numbers, dirty signals, automation vs judgment, comparisons, culture).\n- Consistent “next step” prompt after each playbook: return to topics or hand off to a human.\n- Single escalation point to a defined department (Data Operations) with a clear handoff message.\n\n## Step-by-step\n1. **Input (New message)** starts the workflow when someone asks a question in Calypso Inbox.\n2. **Knowledge Base policy** attempts to answer using your maintained guidance (definitions, metric rules, branch event taxonomy, attribution rules, etc.).\n3. If a direct answer isn’t available, the workflow shows **“Choose what you’re deciding”** with five buttons:\n   1) *Which branch numbers deserve trust (vs noise)*\n   2) *Spot dirty signal before the meeting* \n   3) *Automation vs human judgment* \n   4) *Comparing branches & attribution traps* \n   5) *Build a signal culture (decisions > slides)*\n4. The workflow routes based on the selected button and sends a focused checklist as a **text message**.\n5. After the checklist, **“What do you want to do next?”** appears:\n   - **Back to topics** returns to the main menu.\n   - **Talk to Data Ops** escalates to the **Data Operations** department via fallback handoff.\n\n## Setup requirements\n- **Calypso Knowledge Base**: Add/maintain entries for metric definitions, branch comparability rules, attribution logic, and known caveats. No external credentials required.\n- **Calypso Inbox channel that supports interactive buttons** (where you deploy this workflow).\n- **A configured handoff department** named **Data Operations** (match the department ID/name used in the fallback node).",{"id":13,"teamId":14,"name":9,"version":15,"workflowVersion":16,"nodes":17,"connections":194,"routingEnabled":8,"active":33},"wf_signal_reliability_navigator_v1","calypso-public-library","1.0.0",1,[18,34,40,52,79,88,96,102,109,115,122,128,135,141,148,163,170,181,187],{"id":19,"name":20,"type":21,"typeVersion":16,"position":22,"parameters":25,"category":32,"deletable":33,"connectable":33},"cfg_1","Workflow settings","flow-configs",[23,24],80,40,{"name":9,"description":26,"tags":27,"triggerType":31},"Guided triage for branch numbers, dirty signals, automation judgment calls, comparisons, and building a signal culture.",[28,29,30],"signal-quality","branch-metrics","decision-making","input","policy",false,{"id":35,"name":36,"type":31,"typeVersion":16,"position":37,"parameters":39,"category":31,"deletable":33,"connectable":8},"input_1","Inbound message",[23,38],140,{},{"id":41,"name":42,"type":43,"typeVersion":16,"position":44,"parameters":46,"category":51,"deletable":8,"connectable":8},"kb_1","Knowledge Base answer (when clear)","knowledge-base-policy",[45,38],320,{"enabled":8,"fallbackToRouting":8,"sticky":33,"stickyMode":47,"activationOpener":48,"personalization":50},"default",{"enabled":33,"instruction":49},"",{"useContactName":33},"response",{"id":53,"name":54,"type":55,"typeVersion":16,"position":56,"parameters":58,"category":51,"deletable":8,"connectable":8},"menu_1","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","Decision menu","Pick the check you need. Built to catch clean-looking bad data before it wins a meeting.","Tip: Ask a metric name for a fast answer.",[64,67,70,73,76],{"id":65,"title":66},"trust_numbers","Trust numbers",{"id":68,"title":69},"spot_dirty_signal","Dirty signal",{"id":71,"title":72},"auto_vs_judgment","Auto vs judgment",{"id":74,"title":75},"compare_branches","Compare branches",{"id":77,"title":78},"signal_culture","Signal culture",{"id":80,"name":81,"type":82,"typeVersion":16,"position":83,"parameters":85,"category":87,"deletable":8,"connectable":8},"if_1","If: Trustworthy numbers vs noise","if",[84,38],800,{"buttonId":65,"operator":86},"equals","routing",{"id":89,"name":90,"type":91,"typeVersion":16,"position":92,"parameters":94,"category":51,"deletable":8,"connectable":8},"txt_trust_1","Playbook: Which numbers deserve trust","text-message",[93,24],1040,{"text":95},"Trust the branch number *only after* it passes these sanity checks:\n\n1) **Definition stability**: Same metric definition this month as last month? If the definition moved, the trend is a costume.\n2) **Collection consistency**: Same source systems, same time window, same cutoffs across branches. “Looks aligned” isn’t aligned.\n3) **Completeness & latency**: What % arrives late or gets backfilled? Numbers that ‘settle’ after the meeting are not decision-safe.\n4) **Incentive pressure**: If someone’s bonus loves the metric, assume it’s being gamed until proven otherwise.\n5) **Counter-signal check**: One independent cross-check (ops events, conversation tags, exception logs). If nothing corroborates it, treat it as a hypothesis, not a fact.\n\nRule of thumb: **A metric is trustworthy when it can disagree with you without breaking.**",{"id":97,"name":98,"type":82,"typeVersion":16,"position":99,"parameters":101,"category":87,"deletable":8,"connectable":8},"if_2","If: Spot dirty signal fast",[84,100],260,{"buttonId":68,"operator":86},{"id":103,"name":104,"type":91,"typeVersion":16,"position":105,"parameters":107,"category":51,"deletable":8,"connectable":8},"txt_dirty_1","Playbook: Spot dirty signal before it hurts",[93,106],160,{"text":108},"Dirty signal is polite: it rarely announces itself. Look for these early warnings:\n\n- **Too-smooth charts**: Real branches are messy. Perfect lines often mean filtering, bucketing, or missing data.\n- **One branch is ‘special’**: Different process, different staffing, different system version. Comparisons become theater.\n- **Sudden step-changes**: If the world didn’t change, your instrumentation did.\n- **High confidence, low curiosity**: People stop asking “how is this measured?” right before they regret it.\n- **Conversation mismatch**: Frontline chats sound worse/better than the KPI. That gap is where the truth is hiding.\n\nMeeting save tactic: ask one question out loud—**“What would make this number wrong?”** If no one can answer, pause the decision.",{"id":110,"name":111,"type":82,"typeVersion":16,"position":112,"parameters":114,"category":87,"deletable":8,"connectable":8},"if_3","If: Automation vs judgment",[84,113],380,{"buttonId":71,"operator":86},{"id":116,"name":117,"type":91,"typeVersion":16,"position":118,"parameters":120,"category":51,"deletable":8,"connectable":8},"txt_auto_1","Playbook: When to trust automation",[93,119],280,{"text":121},"Trust automation when the decision is **repeatable**, the inputs are **well-instrumented**, and the cost of a wrong call is **bounded**.\n\nUse automation for:\n- Routine triage (known patterns, stable thresholds)\n- Alerts with clear runbooks\n- Low-stakes prioritization\n\nRequire human judgment when:\n- **The label can be gamed** (people can change behavior to ‘look good’)\n- **The context is novel** (policy change, product change, staffing change)\n- **Errors are asymmetric** (one wrong decision is very expensive)\n- **The metric is a proxy** (attribution and intent are especially slippery)\n\nPractical rule: If you can’t explain the logic in two sentences to a skeptical branch manager, don’t let it auto-decide. Let it *recommend*.",{"id":123,"name":124,"type":82,"typeVersion":16,"position":125,"parameters":127,"category":87,"deletable":8,"connectable":8},"if_4","If: Compare branches & attribution",[84,126],500,{"buttonId":74,"operator":86},{"id":129,"name":130,"type":91,"typeVersion":16,"position":131,"parameters":133,"category":51,"deletable":8,"connectable":8},"txt_compare_1","Playbook: Comparisons & attribution traps",[93,132],400,{"text":134},"Branch comparisons fail in the same predictable ways:\n\n1) **Different denominators**: Headcount, footfall, catchment mix, business hours—normalize or don’t compare.\n2) **Selection effects**: One branch gets tougher cases (or easier ones). Performance may be case mix, not execution.\n3) **Attribution optimism**: If a conversation happened, we credit it. If it didn’t, we ignore it. That’s how you ‘prove’ anything.\n4) **Unequal tooling**: Different CRM usage, tagging discipline, or scripts create measurement differences disguised as outcomes.\n5) **Timing mismatch**: Marketing and ops effects have lags; branches do not respond on the same schedule.\n\nDecision-safe comparison: pick *one* outcome metric + *one* process metric, and ensure both are measured the same way across branches.",{"id":136,"name":137,"type":82,"typeVersion":16,"position":138,"parameters":140,"category":87,"deletable":8,"connectable":8},"if_5","If: Build a signal culture",[84,139],620,{"buttonId":77,"operator":86},{"id":142,"name":143,"type":91,"typeVersion":16,"position":144,"parameters":146,"category":51,"deletable":8,"connectable":8},"txt_culture_1","Playbook: Build a signal culture (decisions > slides)",[93,145],520,{"text":147},"A healthy signal culture doesn’t produce more dashboards. It produces faster, cleaner decisions.\n\nBuild it with these habits:\n- **Name the decision first** (“What will we do differently if this is true?”). No decision = no metric.\n- **Define ‘good enough’ evidence** (what level of uncertainty is acceptable).\n- **Reward surfacing caveats** (the person who finds the measurement flaw is saving you money).\n- **Keep a ‘known issues’ list** for metrics (latency, backfills, branch exceptions). Make it visible.\n- **Close the loop**: after a decision, check whether the signal predicted reality. If it didn’t, fix the signal—not the story.\n\nWitty but true: If your numbers always agree with leadership’s intuition, you don’t have analytics—you have a mirror.",{"id":149,"name":150,"type":55,"typeVersion":16,"position":151,"parameters":153,"category":51,"deletable":8,"connectable":8},"next_1","What next?",[152,119],1280,{"messageType":59,"headerText":154,"bodyText":155,"footerText":49,"sectionTitle":49,"buttons":156,"ctaDisplayText":49,"ctaUrl":49},"Next step","Want to ask something else, or should we loop in a human?",[157,160],{"id":158,"title":159},"new_question","New question",{"id":161,"title":162},"talk_to_data_ops","Talk to Data Ops",{"id":164,"name":165,"type":82,"typeVersion":16,"position":166,"parameters":169,"category":87,"deletable":8,"connectable":8},"if_next_talk","If: Talk to Data Ops",[167,168],1520,360,{"buttonId":161,"operator":86},{"id":171,"name":172,"type":173,"typeVersion":16,"position":174,"parameters":176,"category":180,"deletable":8,"connectable":8},"fallback_1","Handoff to Data Operations","fallback",[175,168],1760,{"handoffMessage":177,"departmentId":178,"departmentName":179},"Looping in Data Operations to help validate signals, definitions, and cross-branch comparability. Please share: metric name(s), time range, branches involved, and what decision you’re trying to make.","dept_data_ops","Data Operations","terminal",{"id":182,"name":183,"type":91,"typeVersion":16,"position":184,"parameters":185,"category":51,"deletable":8,"connectable":8},"txt_restart_1","Start a new question",[175,126],{"text":186},"Send a new message anytime to start over. If you include the metric name, branches, and time range, you’ll get a faster (and safer) answer.",{"id":188,"name":189,"type":173,"typeVersion":16,"position":190,"parameters":192,"category":180,"deletable":8,"connectable":8},"fallback_2","Fallback: No matching choice",[93,191],700,{"handoffMessage":193,"departmentId":178,"departmentName":179},"I couldn’t match that choice to a topic. I’ll route you to Data Operations—share the metric, branches, and the decision you’re facing.",[195,199,201,203,206,209,211,213,215,217,219,221,223,225,227,229,231,233,235,237,239],{"id":196,"source":35,"target":41,"sourceHandle":197,"targetHandle":198,"type":47},"c_input_to_kb","out","in",{"id":200,"source":41,"target":53,"sourceHandle":173,"targetHandle":198,"type":47},"c_kb_to_menu",{"id":202,"source":53,"target":80,"sourceHandle":197,"targetHandle":198,"type":47},"c_menu_to_if1",{"id":204,"source":80,"target":89,"sourceHandle":205,"targetHandle":198,"type":47},"c_if1_true_to_txt_trust","true",{"id":207,"source":80,"target":97,"sourceHandle":208,"targetHandle":198,"type":47},"c_if1_false_to_if2","false",{"id":210,"source":97,"target":103,"sourceHandle":205,"targetHandle":198,"type":47},"c_if2_true_to_txt_dirty",{"id":212,"source":97,"target":110,"sourceHandle":208,"targetHandle":198,"type":47},"c_if2_false_to_if3",{"id":214,"source":110,"target":116,"sourceHandle":205,"targetHandle":198,"type":47},"c_if3_true_to_txt_auto",{"id":216,"source":110,"target":123,"sourceHandle":208,"targetHandle":198,"type":47},"c_if3_false_to_if4",{"id":218,"source":123,"target":129,"sourceHandle":205,"targetHandle":198,"type":47},"c_if4_true_to_txt_compare",{"id":220,"source":123,"target":136,"sourceHandle":208,"targetHandle":198,"type":47},"c_if4_false_to_if5",{"id":222,"source":136,"target":142,"sourceHandle":205,"targetHandle":198,"type":47},"c_if5_true_to_txt_culture",{"id":224,"source":136,"target":188,"sourceHandle":208,"targetHandle":198,"type":47},"c_if5_false_to_fallback2",{"id":226,"source":89,"target":149,"sourceHandle":197,"targetHandle":198,"type":47},"c_txt_trust_to_next",{"id":228,"source":103,"target":149,"sourceHandle":197,"targetHandle":198,"type":47},"c_txt_dirty_to_next",{"id":230,"source":116,"target":149,"sourceHandle":197,"targetHandle":198,"type":47},"c_txt_auto_to_next",{"id":232,"source":129,"target":149,"sourceHandle":197,"targetHandle":198,"type":47},"c_txt_compare_to_next",{"id":234,"source":142,"target":149,"sourceHandle":197,"targetHandle":198,"type":47},"c_txt_culture_to_next",{"id":236,"source":149,"target":164,"sourceHandle":197,"targetHandle":198,"type":47},"c_next_to_if_talk",{"id":238,"source":164,"target":171,"sourceHandle":205,"targetHandle":198,"type":47},"c_if_talk_true_to_fallback",{"id":240,"source":164,"target":182,"sourceHandle":208,"targetHandle":198,"type":47},"c_if_talk_false_to_restart","automation",[28,29,30,243,244,245],"data-hygiene","attribution","triage",[247],"Calypso Inbox","intermediate","Calypso","2026-06-22T11:04:11.177Z","/en/workflows/signal-reliability-navigator-for-branch-decisions",{"en":251},{"title":9,"description":254,"ogDescription":255,"twitterDescription":256,"canonicalPath":251,"robots":257,"schemaType":258,"alternates":259},"Guide teams to trust the right branch signals, spot dirty data early, and decide when automation needs human judgment—via a routed chat flow.","A routed chat workflow that turns messy branch metrics and conversations into decision ready checks: trust signals, catch dirty data, and escalate when needed.","Stop clean looking bad data from winning meetings. Route users through practical signal checks and escalate to Data Ops when judgment is required.","index,follow","HowTo",[260],{"hreflang":6,"href":251},1785947667836]