[{"data":1,"prerenderedAt":274},["ShallowReactive",2],{"/en/workflows/branch-signal-research-decision-coach":3},{"id":4,"slug":5,"locale":6,"translationGroupId":7,"localeSwitchApproved":8,"title":9,"description":10,"documentationMarkdown":11,"workflowJson":12,"category":257,"tags":258,"integrations":259,"difficulty":261,"author":262,"verified":36,"featured":36,"date":263,"modified":263,"icon":7,"imageSrc":7,"path":264,"alternates":265,"seo":266},"13b69ccf-73d0-4215-84f6-714607f703be","branch-signal-research-decision-coach","en",null,true,"Branch Signal Research & Decision Coach","An advisor-style decision coach that helps teams separate trustworthy branch numbers from polished noise, catch dirty signals early, and decide when automation is safe vs when judgment should stay in the loop.","## How it works\nThis workflow turns “we have numbers” into “we have *decision‑grade* numbers.” It routes a teammate through a short set of practical checks—trust vs noise, dirty signal detection, automation guardrails, branch comparisons, and signal culture—without forcing them into academic frameworks.\n\nIt’s built for the moment right before a confident meeting goes off the rails: when the dashboard looks clean, the story is compelling, and the wrong conclusion is seconds away. Users pick what they’re deciding, get a crisp set of checks and prompts, then either run another check or hand off to a human for deeper review.\n\n## Key features\n- Starts with a Knowledge Base policy so free‑form questions can be answered, then routes users via a button menu.\n- Five decision-shaped playbooks: trusted branch numbers, dirty-signal spotting, automation vs judgment, branch comparisons, and building signal culture.\n- A “run another check” loop so teams can move from messy evidence to usable insight without losing context.\n- A built-in escalation path to a human team (fallback/handoff) when the signal is too risky or political to automate.\n\n## Step-by-step\n1. **User message triggers the workflow** (Input).\n2. **Knowledge Base policy runs first** to answer direct questions when possible, then continues to routing.\n3. **User chooses what they need** from a button menu:\n   - Which branch numbers deserve trust\n   - Spot dirty signal before the meeting\n   - Automation vs human judgment\n   - Compare branches & attribution\n   - Build a signal culture (decisions, not slides)\n   - Talk to a human\n4. **Workflow routes to the matching playbook** and sends a short, operator-friendly checklist and guidance.\n5. **User chooses the next step**:\n   - **Run another check** (returns to the main menu)\n   - **Talk to a human** (handoff via fallback)\n\n## Setup requirements\n- **Calypso Inbox** enabled for the channel(s) where users will message this workflow.\n- **Optional:** Configure the fallback/handoff destination (department) so “Talk to a human” reaches the right team.\n- No external credentials are required by this workflow.",{"id":13,"teamId":14,"name":9,"version":15,"workflowVersion":16,"nodes":17,"connections":201,"routingEnabled":8,"active":36},"wf_branch_signal_research_decision_coach","calypso-public-library","1.0.0",1,[18,37,43,55,86,95,100,106,111,117,123,131,137,143,149,155,170,177,183,191],{"id":19,"name":20,"type":21,"typeVersion":16,"position":22,"parameters":25,"category":35,"deletable":36,"connectable":36},"node_flow_configs","Workflow settings","flow-configs",[23,24],80,40,{"name":9,"description":26,"tags":27,"triggerType":34},"Guided checks to turn branch signals into decision-grade conclusions: trust vs noise, dirty signal, automation guardrails, comparisons, and culture.",[28,29,30,31,32,33],"signal-design","data-quality","decision-systems","branch-kpis","attribution","automation-guardrails","input","policy",false,{"id":38,"name":39,"type":34,"typeVersion":16,"position":40,"parameters":42,"category":34,"deletable":36,"connectable":8},"node_input","Inbound message",[23,41],160,{},{"id":44,"name":45,"type":46,"typeVersion":16,"position":47,"parameters":49,"category":54,"deletable":8,"connectable":8},"node_kb_policy","Knowledge Base guardrails","knowledge-base-policy",[48,41],320,{"enabled":8,"fallbackToRouting":8,"sticky":8,"stickyMode":50,"activationOpener":51,"personalization":53},"sticky",{"enabled":8,"instruction":52},"I can help you sanity-check branch signals and decision stories. If you have a specific question, ask it—or pick a checklist below.",{"useContactName":8},"response",{"id":56,"name":57,"type":58,"typeVersion":16,"position":59,"parameters":61,"category":54,"deletable":8,"connectable":8},"node_main_menu","Decision menu","interactive-message",[60,41],600,{"messageType":62,"headerText":63,"bodyText":64,"footerText":65,"sectionTitle":66,"buttons":67,"ctaDisplayText":66,"ctaUrl":66},"button","Research, signal design & decision systems","Pick what you’re deciding. I’ll give you a tight checklist—designed to catch the kind of “looks fine” data that later becomes an expensive confidence problem.","Tip: Run the dirty-signal check before presenting.","",[68,71,74,77,80,83],{"id":69,"title":70},"trust_branch_numbers","Trust numbers",{"id":72,"title":73},"spot_dirty_signal","Spot dirty signal",{"id":75,"title":76},"automation_vs_judgment","Automation vs human",{"id":78,"title":79},"compare_branches_attribution","Compare branches",{"id":81,"title":82},"build_signal_culture","Build signal culture",{"id":84,"title":85},"talk_to_human","Talk to a human",{"id":87,"name":88,"type":89,"typeVersion":16,"position":90,"parameters":92,"category":94,"deletable":8,"connectable":8},"node_if_trust","If: Trust the branch numbers","if",[91,23],860,{"buttonId":69,"operator":93},"equals","routing",{"id":96,"name":97,"type":89,"typeVersion":16,"position":98,"parameters":99,"category":94,"deletable":8,"connectable":8},"node_if_dirty","If: Spot dirty signal",[91,41],{"buttonId":72,"operator":93},{"id":101,"name":102,"type":89,"typeVersion":16,"position":103,"parameters":105,"category":94,"deletable":8,"connectable":8},"node_if_auto","If: Automation vs judgment",[91,104],240,{"buttonId":75,"operator":93},{"id":107,"name":108,"type":89,"typeVersion":16,"position":109,"parameters":110,"category":94,"deletable":8,"connectable":8},"node_if_compare","If: Compare branches & attribution",[91,48],{"buttonId":78,"operator":93},{"id":112,"name":113,"type":89,"typeVersion":16,"position":114,"parameters":116,"category":94,"deletable":8,"connectable":8},"node_if_culture","If: Build signal culture",[91,115],400,{"buttonId":81,"operator":93},{"id":118,"name":119,"type":89,"typeVersion":16,"position":120,"parameters":122,"category":94,"deletable":8,"connectable":8},"node_if_human","If: Talk to a human",[91,121],480,{"buttonId":84,"operator":93},{"id":124,"name":125,"type":126,"typeVersion":16,"position":127,"parameters":129,"category":54,"deletable":8,"connectable":8},"node_txt_trust","Playbook: Trustworthy branch numbers","text-message",[128,23],1140,{"text":130},"Here’s how to tell *decision-grade* branch numbers from polished noise:\n\n1) **Start with the “boring” checks** (they catch most lies):\n- Same definition this week vs last week? Same cutoff time? Same inclusion rules?\n- Any sudden changes in missing values, retries, or “unknown” buckets?\n\n2) **Trust numbers that have friction**:\n- They reconcile to something else (cash, inventory, settlements, customer confirmations).\n- They have a known error rate, not just a point estimate.\n\n3) **Be suspicious of numbers that are too smooth**:\n- Perfectly monotonic improvements, especially after incentives.\n- Big gains with no operational story (headcount, hours, supply, demand) to match.\n\n4) **Make the meeting safe**:\nAsk: “What would make this number wrong?” If nobody can answer, you’re not measuring—you’re narrating.\n\nIf you want, tell me the metric name + the decision it’s driving, and I’ll suggest the top 3 failure modes to check.",{"id":132,"name":133,"type":126,"typeVersion":16,"position":134,"parameters":135,"category":54,"deletable":8,"connectable":8},"node_txt_dirty","Playbook: Dirty signal detection",[128,41],{"text":136},"Dirty signal usually looks *normal* until you compare it the right way. Run this before you present anything:\n\n**Fast dirty-signal checklist**\n- **Edge spikes:** Are changes concentrated at shift start/end, end-of-month, or right after a policy change?\n- **Data “weather”:** Any pipeline delays, backfills, duplicate events, or schema changes this period?\n- **Silent denominator swaps:** Did the eligible population change (new products, new booking rules, new customer mix)?\n- **Winner’s curse:** Are you looking only at top branches/agents/campaigns after slicing? That’s selection bias wearing a suit.\n- **Conversation artifacts:** A new script/tool can increase “logged outcomes” without increasing real outcomes.\n\n**Two questions that save careers**\n1) “What are we *not* counting anymore?”\n2) “If this is real, what else should move—and did it?”\n\nIf you share what changed (tool/process/incentive), I’ll tell you where dirty signal tends to sneak in first.",{"id":138,"name":139,"type":126,"typeVersion":16,"position":140,"parameters":141,"category":54,"deletable":8,"connectable":8},"node_txt_auto","Playbook: Automation vs human judgment",[128,104],{"text":142},"When should leaders trust automation—and when should they keep a human in the loop?\n\n**Automate when:**\n- The decision is **reversible** (you can undo it cheaply).\n- The signal is **stable** (definitions don’t drift weekly).\n- You can measure **false positives/negatives** without heroics.\n- The action is **bounded** (small levers, limited blast radius).\n\n**Require human judgment when:**\n- The decision is **high-stakes or political** (people will game it).\n- The signal can be **manipulated** by the people measured.\n- You’re using proxy metrics (speed, touches, “intent”) as if they’re outcomes.\n- The environment is changing fast (new product, new pricing, new staffing).\n\n**A practical guardrail:**\nStart with “automation suggests, human approves” until you’ve logged enough exceptions to know what ‘normal failure’ looks like.\n\nTell me what you want to automate (and the cost of a wrong decision). I’ll recommend a safe rollout pattern.",{"id":144,"name":145,"type":126,"typeVersion":16,"position":146,"parameters":147,"category":54,"deletable":8,"connectable":8},"node_txt_compare","Playbook: Comparing branches & attribution",[128,48],{"text":148},"Branch comparisons and attribution are where teams confidently misread reality. Common traps—and fixes:\n\n**1) Composition effects (the #1 offender):**\nIf Branch A ‘improves,’ ask: did the *customers/cases* change, or did performance change?\n- Fix: compare like-for-like cohorts (same product, segment, complexity).\n\n**2) Different opportunity sets:**\nA branch can’t convert what it never receives.\n- Fix: normalize by eligible volume, not total volume.\n\n**3) Conversation metrics are slippery:**\nMore contacts, more notes, more tags can mean better logging—not better outcomes.\n- Fix: anchor to outcomes that reconcile elsewhere.\n\n**4) Attribution tells a story, not the truth:**\nLast-touch is easy; it’s also how you reward whoever arrives at the end.\n- Fix: treat attribution as a *sensitivity analysis*: “Does the decision hold under two reasonable models?”\n\n**Decision rule that keeps you honest:**\nDon’t rank branches unless you can explain the top 2 drivers *and* the top 2 confounders.\n\nIf you share the metric + how branches differ (traffic mix, staffing, hours), I’ll suggest the fairest comparison cut.",{"id":150,"name":151,"type":126,"typeVersion":16,"position":152,"parameters":153,"category":54,"deletable":8,"connectable":8},"node_txt_culture","Playbook: Signal culture (decisions, not slides)",[128,115],{"text":154},"A strong signal culture doesn’t create more dashboards—it creates faster, safer decisions.\n\n**Behaviors to install (lightweight, high impact):**\n- **Every metric gets a “how it fails” note.** Two sentences. Updated when reality breaks it.\n- **One owner per definition.** Commit to “same words mean same thing” across branches.\n- **Decision memos beat slide decks.** Write the decision, evidence, counter‑evidence, and what would change your mind.\n- **Measure the decision, not the story.** Log: what did we decide, based on what, and what happened after.\n\n**A little wit (and truth):**\nIf a metric is used for praise or punishment, assume it will be optimized—and stop pretending it’s neutral.\n\nIf you tell me what decision keeps getting stuck (pricing, staffing, coaching, process), I’ll suggest the smallest set of signals that unblock it.",{"id":156,"name":157,"type":58,"typeVersion":16,"position":158,"parameters":160,"category":54,"deletable":8,"connectable":8},"node_next_steps","Next steps",[159,104],1400,{"messageType":62,"headerText":161,"bodyText":162,"footerText":163,"sectionTitle":66,"buttons":164,"ctaDisplayText":66,"ctaUrl":66},"What next?","Want to run another check, or escalate this to a human review?","High-stakes? Escalate—it’s a feature.",[165,168],{"id":166,"title":167},"back_to_menu","Run another check",{"id":169,"title":85},"escalate_human",{"id":171,"name":172,"type":89,"typeVersion":16,"position":173,"parameters":176,"category":94,"deletable":8,"connectable":8},"node_if_back","If: Run another check",[174,175],1660,220,{"buttonId":166,"operator":93},{"id":178,"name":179,"type":89,"typeVersion":16,"position":180,"parameters":182,"category":94,"deletable":8,"connectable":8},"node_if_escalate","If: Escalate to human",[174,181],300,{"buttonId":169,"operator":93},{"id":184,"name":185,"type":126,"typeVersion":16,"position":186,"parameters":189,"category":54,"deletable":8,"connectable":8},"node_txt_restart","Restart instructions",[187,188],1920,200,{"text":190},"To run another check, just send a new message here and I’ll show the decision menu again. (Yes, it’s slightly manual. No, it won’t lie to you like a “helpful” dashboard.)",{"id":192,"name":193,"type":194,"typeVersion":16,"position":195,"parameters":196,"category":200,"deletable":8,"connectable":8},"node_fallback_handoff","Handoff to Insights Desk","fallback",[187,48],{"handoffMessage":197,"departmentId":198,"departmentName":199},"Got it. I’m handing this to the Insights Desk for a human sanity check. To speed it up, send: (1) the metric name, (2) the decision you’re about to make, (3) the time window, (4) what changed recently.","dept-insights-desk","Insights Desk","terminal",[202,207,209,211,214,217,219,221,223,225,227,229,231,233,235,237,239,241,243,245,247,249,251,253,255],{"id":203,"source":38,"target":44,"sourceHandle":204,"targetHandle":205,"type":206},"conn_input_to_kb","out","in","default",{"id":208,"source":44,"target":56,"sourceHandle":204,"targetHandle":205,"type":206},"conn_kb_to_menu",{"id":210,"source":56,"target":87,"sourceHandle":204,"targetHandle":205,"type":206},"conn_menu_to_if_trust",{"id":212,"source":87,"target":124,"sourceHandle":213,"targetHandle":205,"type":206},"conn_if_trust_true_to_txt","true",{"id":215,"source":87,"target":96,"sourceHandle":216,"targetHandle":205,"type":206},"conn_if_trust_false_to_if_dirty","false",{"id":218,"source":96,"target":132,"sourceHandle":213,"targetHandle":205,"type":206},"conn_if_dirty_true_to_txt",{"id":220,"source":96,"target":101,"sourceHandle":216,"targetHandle":205,"type":206},"conn_if_dirty_false_to_if_auto",{"id":222,"source":101,"target":138,"sourceHandle":213,"targetHandle":205,"type":206},"conn_if_auto_true_to_txt",{"id":224,"source":101,"target":107,"sourceHandle":216,"targetHandle":205,"type":206},"conn_if_auto_false_to_if_compare",{"id":226,"source":107,"target":144,"sourceHandle":213,"targetHandle":205,"type":206},"conn_if_compare_true_to_txt",{"id":228,"source":107,"target":112,"sourceHandle":216,"targetHandle":205,"type":206},"conn_if_compare_false_to_if_culture",{"id":230,"source":112,"target":150,"sourceHandle":213,"targetHandle":205,"type":206},"conn_if_culture_true_to_txt",{"id":232,"source":112,"target":118,"sourceHandle":216,"targetHandle":205,"type":206},"conn_if_culture_false_to_if_human",{"id":234,"source":118,"target":192,"sourceHandle":213,"targetHandle":205,"type":206},"conn_if_human_true_to_handoff",{"id":236,"source":118,"target":192,"sourceHandle":216,"targetHandle":205,"type":206},"conn_if_human_false_to_handoff",{"id":238,"source":124,"target":156,"sourceHandle":204,"targetHandle":205,"type":206},"conn_txt_trust_to_next",{"id":240,"source":132,"target":156,"sourceHandle":204,"targetHandle":205,"type":206},"conn_txt_dirty_to_next",{"id":242,"source":138,"target":156,"sourceHandle":204,"targetHandle":205,"type":206},"conn_txt_auto_to_next",{"id":244,"source":144,"target":156,"sourceHandle":204,"targetHandle":205,"type":206},"conn_txt_compare_to_next",{"id":246,"source":150,"target":156,"sourceHandle":204,"targetHandle":205,"type":206},"conn_txt_culture_to_next",{"id":248,"source":156,"target":171,"sourceHandle":204,"targetHandle":205,"type":206},"conn_next_to_if_back",{"id":250,"source":171,"target":184,"sourceHandle":213,"targetHandle":205,"type":206},"conn_if_back_true_to_restart",{"id":252,"source":171,"target":178,"sourceHandle":216,"targetHandle":205,"type":206},"conn_if_back_false_to_if_escalate",{"id":254,"source":178,"target":192,"sourceHandle":213,"targetHandle":205,"type":206},"conn_if_escalate_true_to_handoff",{"id":256,"source":178,"target":184,"sourceHandle":216,"targetHandle":205,"type":206},"conn_if_escalate_false_to_restart","automation",[28,29,30,31,32,33],[260],"Calypso Inbox","intermediate","Calypso","2026-07-06T11:04:42.300Z","/en/workflows/branch-signal-research-decision-coach",{"en":264},{"title":9,"description":267,"ogDescription":268,"twitterDescription":269,"canonicalPath":264,"robots":270,"schemaType":271,"alternates":272},"Help teams trust the right branch numbers, spot dirty signal early, and choose when to automate vs use judgment—via guided checks and handoff.","A practical decision coach for branch metrics: separate trustworthy numbers from noise, catch dirty signal early, and know when automation is safe vs risky.","Keep branch decisions honest: quick checks for trustworthy numbers, dirty signal detection, automation guardrails, and a clean handoff when judgment is needed.","index,follow","HowTo",[273],{"hreflang":6,"href":264},1785947667568]