{ "schemaVersion": "1.0", "item": { "slug": "afrexai-automation-strategy", "name": "Business Automation Strategy", "source": "tencent", "type": "skill", "category": "开发工具", "sourceUrl": "https://clawhub.ai/1kalin/afrexai-automation-strategy", "canonicalUrl": "https://clawhub.ai/1kalin/afrexai-automation-strategy", "targetPlatform": "OpenClaw" }, "install": { "downloadMode": "redirect", "downloadUrl": "/downloads/afrexai-automation-strategy", "sourceDownloadUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=afrexai-automation-strategy", "sourcePlatform": "tencent", "targetPlatform": "OpenClaw", "installMethod": "Manual import", "extraction": "Extract archive", "prerequisites": [ "OpenClaw" ], "packageFormat": "ZIP package", "includedAssets": [ "README.md", "SKILL.md" ], "primaryDoc": "SKILL.md", "quickSetup": [ "Download the package from Yavira.", "Extract the archive and review SKILL.md first.", "Import or place the package into your OpenClaw setup." ], "agentAssist": { "summary": "Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.", "steps": [ "Download the package from Yavira.", "Extract it into a folder your agent can access.", "Paste one of the prompts below and point your agent at the extracted folder." ], "prompts": [ { "label": "New install", "body": "I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Then review README.md for any prerequisites, environment setup, or post-install checks. Tell me what you changed and call out any manual steps you could not complete." }, { "label": "Upgrade existing", "body": "I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Then review README.md for any prerequisites, environment setup, or post-install checks. Summarize what changed and any follow-up checks I should run." } ] }, "sourceHealth": { "source": "tencent", "slug": "afrexai-automation-strategy", "status": "healthy", "reason": "direct_download_ok", "recommendedAction": "download", "checkedAt": "2026-05-01T04:31:43.185Z", "expiresAt": "2026-05-08T04:31:43.185Z", "httpStatus": 200, "finalUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=afrexai-automation-strategy", "contentType": "application/zip", "probeMethod": "head", "details": { "probeUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=afrexai-automation-strategy", "contentDisposition": "attachment; filename=\"afrexai-automation-strategy-1.0.0.zip\"", "redirectLocation": null, "bodySnippet": null, "slug": "afrexai-automation-strategy" }, "scope": "item", "summary": "Item download looks usable.", "detail": "Yavira can redirect you to the upstream package for this item.", "primaryActionLabel": "Download for OpenClaw", "primaryActionHref": "/downloads/afrexai-automation-strategy" }, "validation": { "installChecklist": [ "Use the Yavira download entry.", "Review SKILL.md after the package is downloaded.", "Confirm the extracted package contains the expected setup assets." ], "postInstallChecks": [ "Confirm the extracted package includes the expected docs or setup files.", "Validate the skill or prompts are available in your target agent workspace.", "Capture any manual follow-up steps the agent could not complete." ] }, "downloadPageUrl": "https://openagent3.xyz/downloads/afrexai-automation-strategy", "agentPageUrl": "https://openagent3.xyz/skills/afrexai-automation-strategy/agent", "manifestUrl": "https://openagent3.xyz/skills/afrexai-automation-strategy/agent.json", "briefUrl": "https://openagent3.xyz/skills/afrexai-automation-strategy/agent.md" }, "agentAssist": { "summary": "Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.", "steps": [ "Download the package from Yavira.", "Extract it into a folder your agent can access.", "Paste one of the prompts below and point your agent at the extracted folder." ], "prompts": [ { "label": "New install", "body": "I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Then review README.md for any prerequisites, environment setup, or post-install checks. Tell me what you changed and call out any manual steps you could not complete." }, { "label": "Upgrade existing", "body": "I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Then review README.md for any prerequisites, environment setup, or post-install checks. Summarize what changed and any follow-up checks I should run." } ] }, "documentation": { "source": "clawhub", "primaryDoc": "SKILL.md", "sections": [ { "title": "Business Automation Strategy — AfrexAI", "body": "The complete methodology for identifying, designing, building, and scaling business automations. Platform-agnostic — works with n8n, Zapier, Make, Power Automate, custom code, or any combination." }, { "title": "Phase 1: Automation Audit — Find the Gold", "body": "Before building anything, map where time and money leak." }, { "title": "Quick ROI Triage", "body": "Ask these 5 questions about any process:\n\nHow often does it happen? (frequency)\nHow long does it take? (duration per occurrence)\nHow many people touch it? (handoffs)\nHow error-prone is it? (failure rate)\nHow much does failure cost? (impact)" }, { "title": "Process Inventory Template", "body": "process_inventory:\n process_name: \"[Name]\"\n department: \"[Sales/Marketing/Ops/Finance/HR/Engineering]\"\n owner: \"[Person responsible]\"\n frequency: \"[X per day/week/month]\"\n duration_minutes: [time per occurrence]\n monthly_volume: [total occurrences]\n monthly_hours: [volume × duration ÷ 60]\n hourly_cost: [fully loaded employee cost]\n monthly_cost: \"$[hours × hourly cost]\"\n error_rate: \"[X%]\"\n error_cost_per_incident: \"$[average]\"\n handoffs: [number of people involved]\n current_tools: [\"tool1\", \"tool2\"]\n automation_potential: \"[Full/Partial/Assist/None]\"\n complexity: \"[Simple/Medium/Complex/Enterprise]\"\n dependencies: [\"system1\", \"system2\"]\n notes: \"[Pain points, workarounds, tribal knowledge]\"" }, { "title": "Automation Potential Classification", "body": "LevelDescriptionHuman RoleExampleFullEnd-to-end automated, no human neededMonitor exceptionsInvoice processing, data syncPartialAutomated with human approval gatesReview & approveContract generation, hiring workflowAssistHuman does work, automation helpsExecute with AI assistanceCustomer support, content creationNoneRequires human judgment/creativityFull ownershipStrategy, relationship building" }, { "title": "ROI Calculation", "body": "Annual savings = (monthly_hours × 12 × hourly_cost) + (error_rate × volume × 12 × error_cost)\nBuild cost = development_hours × developer_rate + tool_costs\nPayback period = build_cost ÷ (annual_savings ÷ 12) months\nROI = ((annual_savings - annual_tool_cost) ÷ build_cost) × 100%\n\nDecision rules:\n\nPayback < 3 months → Build immediately\nPayback 3-6 months → Build this quarter\nPayback 6-12 months → Evaluate against alternatives\nPayback > 12 months → Reconsider (unless strategic)" }, { "title": "ICE-R Scoring (0-10 each)", "body": "DimensionWeightScoring GuideImpact30%10=saves >$50K/yr, 7=saves >$20K/yr, 5=saves >$5K/yr, 3=saves >$1K/yrConfidence20%10=proven pattern, 7=similar done before, 5=feasible but new, 3=uncertainEase25%10=<1 day, 7=<1 week, 5=<1 month, 3=<3 months, 1=>3 monthsReliability25%10=deterministic, 7=95%+ success, 5=80%+ success, 3=needs frequent fixes\n\nScore = (Impact × 0.30) + (Confidence × 0.20) + (Ease × 0.25) + (Reliability × 0.25)" }, { "title": "Quick Win Identification", "body": "Automate FIRST (highest ROI, lowest risk):\n\nData entry / copy-paste between systems\nNotification routing (email → Slack → SMS based on rules)\nReport generation and distribution\nFile organization and naming\nStatus updates across tools\nMeeting scheduling and follow-ups\nInvoice creation from templates\nLead capture → CRM entry\nOnboarding checklists\nBackup and archival\n\nAutomate LAST (complex, high risk):\n\nAnything involving money transfers without approval\nCustomer-facing responses without review\nLegal/compliance decisions\nHiring/firing workflows\nSecurity-sensitive operations" }, { "title": "Platform Decision Matrix", "body": "FactorNo-Code (Zapier/Make)Low-Code (n8n/Power Automate)Custom CodeAI AgentBest forSimple integrationsComplex workflowsUnique logicJudgment callsBuild speedHoursDaysWeeksDays-weeksMaintenanceLowMediumHighMediumFlexibilityLimitedHighUnlimitedHighCost at scaleExpensiveModerateCheapVariesError handlingBasicGoodFull controlVariableTeam skill neededBusiness userTechnical BADeveloperAI engineerVendor lock-inHighMediumNoneLow-medium" }, { "title": "Selection Decision Tree", "body": "Is the process deterministic (same input → same output)?\n├── YES: Does it involve >3 systems?\n│ ├── YES: Does it need complex branching logic?\n│ │ ├── YES → Low-code (n8n/Power Automate)\n│ │ └── NO → No-code (Zapier/Make) if budget allows, else n8n\n│ └── NO: Is it performance-critical?\n│ ├── YES → Custom code\n│ └── NO → No-code (simplest wins)\n└── NO: Does it need judgment/reasoning?\n ├── YES: Is the judgment pattern learnable?\n │ ├── YES → AI agent with human review\n │ └── NO → Human-assisted automation\n └── NO → Partial automation with human gates" }, { "title": "Cost Comparison by Scale", "body": "Monthly TasksZapierMaken8n (self-hosted)Custom Code1,000$30$10$5 (hosting)$50+ (hosting)10,000$100$30$5$50+100,000$500+$150$10$50+1,000,000$2,000+$500+$20$100+\n\nRule: If you're spending >$200/mo on Zapier/Make, evaluate self-hosted n8n." }, { "title": "Workflow Blueprint Template", "body": "workflow_blueprint:\n name: \"[Descriptive name]\"\n id: \"WF-[DEPT]-[NUMBER]\"\n version: \"1.0.0\"\n owner: \"[Person]\"\n priority: \"[P0-P3]\"\n \n trigger:\n type: \"[webhook/schedule/event/manual/condition]\"\n source: \"[System or schedule]\"\n conditions: \"[When to fire]\"\n dedup_strategy: \"[How to prevent double-processing]\"\n \n inputs:\n - name: \"[field]\"\n type: \"[string/number/date/object]\"\n required: true\n validation: \"[rules]\"\n source: \"[where it comes from]\"\n \n steps:\n - id: \"step_1\"\n action: \"[verb: fetch/transform/validate/send/create/update/delete]\"\n system: \"[target system]\"\n description: \"[what this step does]\"\n input: \"[from trigger or previous step]\"\n output: \"[what it produces]\"\n error_handling: \"[retry/skip/alert/abort]\"\n timeout_seconds: 30\n \n - id: \"step_2_branch\"\n type: \"condition\"\n condition: \"[expression]\"\n true_path: \"step_3a\"\n false_path: \"step_3b\"\n \n error_handling:\n retry_policy:\n max_attempts: 3\n backoff: \"exponential\"\n initial_delay_seconds: 5\n on_failure: \"[alert/queue-for-review/fallback]\"\n alert_channel: \"[Slack/email/SMS]\"\n dead_letter_queue: true\n \n monitoring:\n success_metric: \"[what defines success]\"\n expected_duration_seconds: [max]\n alert_on_duration_exceeded: true\n log_level: \"[info/debug/error]\"\n \n testing:\n test_data: \"[how to generate test inputs]\"\n expected_output: \"[what success looks like]\"\n edge_cases: [\"empty input\", \"duplicate\", \"malformed data\"]" }, { "title": "7 Workflow Design Principles", "body": "Idempotent by default — Running the same workflow twice with the same input should produce the same result, not duplicates\nFail loudly — Silent failures are worse than crashes. Every error must notify someone\nCheckpoint progress — Long workflows should save state so they can resume, not restart\nValidate early — Check inputs at the start, not after 10 expensive API calls\nSeparate concerns — One workflow, one job. Chain workflows, don't build monoliths\nLog everything — Timestamps, inputs, outputs, decisions. You WILL need to debug\nHuman escape hatch — Every automated workflow needs a manual override path" }, { "title": "Common Workflow Patterns", "body": "PatternWhen to UseExampleSequentialSteps depend on each otherLead → Enrich → Score → RouteParallel fan-outIndependent stepsSend email + Update CRM + Log analyticsConditional branchDifferent paths by dataHigh value → Sales, Low value → NurtureLoop/batchProcess collectionsFor each row in CSV, create recordApproval gateHuman judgment neededContract review before sendingEvent-driven chainWorkflow triggers workflowOrder placed → Fulfillment → Shipping → NotificationRetry with fallbackUnreliable external APIsTry API → Retry 3x → Use cached data → AlertScheduled sweepPeriodic cleanup/syncNightly: sync CRM → accounting" }, { "title": "Integration Quality Checklist", "body": "For every system integration:\n\nAPI documentation reviewed\n Authentication method confirmed (OAuth2/API key/JWT)\n Rate limits documented (requests/min, requests/day)\n Webhook support checked (push vs poll)\n Error response format understood\n Pagination handling planned\n Data format confirmed (JSON/XML/CSV)\n Field mapping documented\n Test environment available\n Sandbox/production separation configured" }, { "title": "Data Mapping Template", "body": "data_mapping:\n source_system: \"[System A]\"\n target_system: \"[System B]\"\n sync_direction: \"[one-way/bidirectional]\"\n sync_frequency: \"[real-time/5min/hourly/daily]\"\n conflict_resolution: \"[source wins/target wins/newest wins/manual]\"\n \n field_mappings:\n - source_field: \"contact.email\"\n target_field: \"customer.email_address\"\n transform: \"lowercase\"\n required: true\n - source_field: \"contact.company\"\n target_field: \"customer.organization\"\n transform: \"trim\"\n default: \"Unknown\"\n - source_field: \"contact.created_at\"\n target_field: \"customer.signup_date\"\n transform: \"ISO8601 → YYYY-MM-DD\"" }, { "title": "Rate Limit Strategy", "body": "ApproachWhenImplementationQueue + throttlePredictable volumeProcess queue at 80% of rate limitExponential backoffBurst trafficWait 1s, 2s, 4s, 8s on 429 errorsBatch API callsHigh volume CRUDGroup 50-100 records per callCache responsesRepeated lookupsCache for TTL matching data freshness needsOff-peak schedulingNon-urgent syncsRun heavy syncs at 2-4 AM" }, { "title": "Error Classification", "body": "TypeExampleResponsePriorityTransientAPI timeout, 503Retry with backoffAuto-handleRate limit429 Too Many RequestsQueue + throttleAuto-handleData validationMissing required fieldLog + skip + alertReview dailyAuth failureToken expiredRefresh + retry, else alertP1 — fix within 1hLogic errorUnexpected stateHalt + alert + queueP0 — fix immediatelyExternal changeAPI schema changedHalt + alertP0 — fix immediatelyCapacityQueue overflowScale + alertP1 — fix within 4h" }, { "title": "Dead Letter Queue Pattern", "body": "Every workflow should have a DLQ:\n\nCapture — Failed items go to DLQ with full context (input, error, timestamp, step)\nAlert — Notify on DLQ growth (>10 items or >1% failure rate)\nReview — Daily check of DLQ items\nReplay — Ability to reprocess DLQ items after fix\nExpire — Auto-archive items older than 30 days with summary" }, { "title": "Circuit Breaker Pattern", "body": "States: CLOSED (normal) → OPEN (failing) → HALF-OPEN (testing)\n\nCLOSED: Process normally, track failures\n → If failure_count > threshold in window → OPEN\n\nOPEN: Reject all requests, return cached/default\n → After cool_down_period → HALF-OPEN\n\nHALF-OPEN: Allow 1 test request\n → If success → CLOSED\n → If failure → OPEN (reset cool_down)\n\nThresholds:\n\nSimple integrations: 5 failures in 60 seconds\nCritical paths: 3 failures in 30 seconds\nNon-critical: 10 failures in 300 seconds" }, { "title": "Automation Test Pyramid", "body": "LevelWhatHowWhenUnitIndividual step logicMock inputs, verify outputEvery changeIntegrationSystem connectionsTest with sandbox APIsWeekly + after changesEnd-to-endFull workflow pathRun with test dataBefore deploy + weeklyChaosFailure scenariosKill steps, corrupt dataMonthlyLoadVolume handling10x normal volumeBefore scaling" }, { "title": "Test Scenario Checklist", "body": "For every workflow, test:\n\nHappy path (normal input, expected output)\n Empty/null input (missing required fields)\n Duplicate input (same event twice)\n Malformed input (wrong types, encoding issues)\n Boundary values (max length, zero, negative)\n API down (target system unavailable)\n Slow response (timeout handling)\n Partial failure (step 3 of 5 fails)\n Concurrent execution (two runs at same time)\n Clock skew / timezone issues\n Large payload (oversized data)\n Permission denied (auth issues)" }, { "title": "Validation Before Go-Live", "body": "go_live_checklist:\n functionality:\n - [ ] All test scenarios pass\n - [ ] Edge cases documented and handled\n - [ ] Error messages are actionable\n \n reliability:\n - [ ] Retry logic tested\n - [ ] Circuit breaker configured\n - [ ] Dead letter queue active\n - [ ] Idempotency verified (run twice, same result)\n \n monitoring:\n - [ ] Success/failure alerts configured\n - [ ] Duration alerts set\n - [ ] Log retention configured\n - [ ] Dashboard created\n \n documentation:\n - [ ] Workflow blueprint updated\n - [ ] Runbook written\n - [ ] Team trained on manual override\n \n rollback:\n - [ ] Previous version preserved\n - [ ] Rollback procedure tested\n - [ ] Data cleanup plan for partial runs" }, { "title": "Automation Health Dashboard", "body": "automation_dashboard:\n period: \"weekly\"\n \n summary:\n total_workflows: [count]\n total_executions: [count]\n success_rate: \"[X%]\"\n avg_duration: \"[X seconds]\"\n errors_this_period: [count]\n time_saved_hours: [calculated]\n cost_saved: \"$[calculated]\"\n \n by_workflow:\n - name: \"[Workflow name]\"\n executions: [count]\n success_rate: \"[X%]\"\n avg_duration: \"[X seconds]\"\n p95_duration: \"[X seconds]\"\n errors: [count]\n error_types: [\"type1: count\", \"type2: count\"]\n dlq_items: [count]\n status: \"[healthy/degraded/failing]\"\n \n alerts_fired: [count]\n manual_interventions: [count]\n \n top_issues:\n - \"[Issue 1: description + fix status]\"\n - \"[Issue 2: description + fix status]\"\n \n cost:\n platform_cost: \"$[monthly]\"\n api_calls_cost: \"$[monthly]\"\n compute_cost: \"$[monthly]\"\n total: \"$[monthly]\"\n cost_per_execution: \"$[calculated]\"" }, { "title": "Alert Rules", "body": "MetricWarningCriticalActionSuccess rate<95%<90%Investigate + fixDuration>2x average>5x averageCheck for bottleneckDLQ size>10 items>50 itemsReview + reprocessError spike5 errors/hour20 errors/hourPause + investigateQueue depth>100 pending>1000 pendingScale + investigateCost spike>150% of average>300% of averageAudit + optimize" }, { "title": "Weekly Review Questions", "body": "Which workflows had the lowest success rate? Why?\nAre any workflows consistently slow? What's the bottleneck?\nHow many manual interventions were needed? Can we eliminate them?\nWhat's in the DLQ? Patterns?\nAre we approaching any rate limits?\nTotal cost vs total time saved — still positive ROI?" }, { "title": "Scaling Checklist", "body": "Before scaling any automation:\n\nLoad tested at 10x current volume\n Rate limits mapped for all APIs\n Queue-based architecture (not synchronous chains)\n Database indexes optimized\n Caching layer in place\n Monitoring alerts adjusted for new thresholds\n Cost projections at scale calculated\n Fallback/degradation plan documented" }, { "title": "Performance Optimization Priority", "body": "Eliminate unnecessary API calls — Cache lookups, batch operations\nParallelize independent steps — Don't wait when you don't have to\nOptimize data payloads — Only fetch/send fields you need\nUse webhooks over polling — Real-time + fewer API calls\nBatch processing — Group operations (50-100 per batch)\nAsync where possible — Don't block on non-critical steps\nCDN/cache for static lookups — Country codes, categories, templates\nDatabase query optimization — Indexes, query plans, connection pooling" }, { "title": "When to Migrate Platforms", "body": "SignalFromToSpending >$500/mo on Zapier/MakeNo-codeSelf-hosted n8nNeed custom logic in >50% of workflowsNo-codeLow-code or code>100K executions/dayAny hostedSelf-hosted or customComplex branching breaking visual toolsLow-codeCustom codeMultiple teams building automationsSingle toolPlatform + governanceAI judgment needed in workflowsTraditionalAI agent integration" }, { "title": "Automation Registry", "body": "Every automation must be registered:\n\nautomation_registry_entry:\n id: \"WF-[DEPT]-[NUMBER]\"\n name: \"[Descriptive name]\"\n description: \"[What it does in one sentence]\"\n owner: \"[Person]\"\n team: \"[Department]\"\n platform: \"[n8n/Zapier/Make/custom]\"\n status: \"[active/paused/deprecated/testing]\"\n created: \"[date]\"\n last_modified: \"[date]\"\n last_reviewed: \"[date]\"\n review_frequency: \"[monthly/quarterly]\"\n \n business_impact:\n time_saved_monthly_hours: [X]\n cost_saved_monthly: \"$[X]\"\n error_reduction: \"[X%]\"\n \n technical:\n trigger: \"[type]\"\n systems_connected: [\"system1\", \"system2\"]\n avg_daily_executions: [X]\n success_rate: \"[X%]\"\n \n dependencies:\n upstream: [\"WF-XXX\"]\n downstream: [\"WF-YYY\"]\n \n documentation:\n blueprint: \"[link]\"\n runbook: \"[link]\"\n test_plan: \"[link]\"" }, { "title": "Naming Conventions", "body": "Pattern: [DEPT]-[ACTION]-[OBJECT]-[QUALIFIER]\nExamples:\n SALES-sync-leads-from-typeform\n FINANCE-generate-invoice-monthly\n HR-onboard-employee-new-hire\n MARKETING-post-content-social-scheduled\n OPS-backup-database-nightly" }, { "title": "Change Management for Automations", "body": "Change TypeApprovalTestingRollback PlanConfig change (threshold, timing)OwnerQuick smoke testRevert configLogic change (new branch, new step)Owner + reviewerFull test suitePrevious versionIntegration change (new API, new system)Owner + tech leadIntegration + E2EDisconnect + manualNew workflowOwner + stakeholderFull test + pilotDisable workflowDeprecationOwner + affected teamsVerify replacementsRe-enable" }, { "title": "Quarterly Automation Review", "body": "Inventory check — Are all automations in the registry? Any rogue workflows?\nROI validation — Is each automation still delivering value?\nHealth review — Success rates, error trends, DLQ patterns\nCost audit — Platform costs trending up? Optimization opportunities?\nSecurity review — API keys rotated? Permissions still appropriate?\nDeprecation candidates — Any automations that should be retired?\nOpportunity scan — New processes to automate? Existing ones to improve?" }, { "title": "When to Add AI to Automations", "body": "ScenarioAI TypeExampleClassify unstructured textLLMCategorize support ticketsExtract data from documentsLLM + OCRParse invoices, contractsGenerate content from templatesLLMPersonalized emails, reportsMake judgment callsLLM + rulesLead scoring, risk assessmentSummarize informationLLMMeeting notes, research briefsRoute based on intentLLMCustomer request → right team" }, { "title": "AI Integration Best Practices", "body": "Always validate AI output — LLMs hallucinate. Add validation checks\nSet confidence thresholds — Below threshold → human review queue\nLog AI decisions — Input, output, confidence, model version\nA/B test AI vs rules — Prove AI adds value before committing\nCost-control AI calls — Cache similar inputs, batch where possible\nFallback to rules — If AI is unavailable, have deterministic backup\nReview AI decisions weekly — Spot check for quality drift" }, { "title": "AI Agent Integration Pattern", "body": "ai_agent_step:\n type: \"ai_judgment\"\n model: \"[model name]\"\n \n input:\n context: \"[relevant data from previous steps]\"\n task: \"[specific instruction — be precise]\"\n output_format: \"[JSON schema or structured format]\"\n constraints: [\"must not\", \"must always\", \"if unsure\"]\n \n validation:\n confidence_threshold: 0.85\n required_fields: [\"field1\", \"field2\"]\n value_ranges:\n score: [0, 100]\n category: [\"A\", \"B\", \"C\"]\n \n on_low_confidence:\n action: \"route_to_human\"\n queue: \"[review queue name]\"\n \n on_failure:\n action: \"fallback_to_rules\"\n rules_engine: \"[rule set name]\"\n \n monitoring:\n log_all_decisions: true\n sample_rate_for_review: 0.10\n alert_on_confidence_drop: true" }, { "title": "5 Levels of Automation Maturity", "body": "LevelNameDescriptionIndicators1Ad HocManual processes, maybe a few scriptsNo registry, tribal knowledge2ReactiveAutomate pain points as they ariseSome workflows, no standards3SystematicPlanned automation programRegistry, testing, monitoring4OptimizedContinuous improvement, governanceROI tracking, quarterly reviews5IntelligentAI-augmented, self-healingAdaptive workflows, predictive" }, { "title": "Maturity Assessment (Score 1-5 per dimension)", "body": "automation_maturity:\n dimensions:\n strategy: [1-5] # Planned roadmap vs ad hoc\n architecture: [1-5] # Patterns, standards, reuse\n reliability: [1-5] # Error handling, monitoring, uptime\n governance: [1-5] # Registry, change management, reviews\n testing: [1-5] # Test coverage, validation, chaos\n documentation: [1-5] # Blueprints, runbooks, training\n optimization: [1-5] # Performance, cost, continuous improvement\n ai_integration: [1-5] # AI-powered decisions, self-healing\n \n total: [sum ÷ 8]\n grade: \"[A/B/C/D/F]\"\n # A: 4.5+ | B: 3.5-4.4 | C: 2.5-3.4 | D: 1.5-2.4 | F: <1.5\n \n top_gap: \"[lowest scoring dimension]\"\n next_action: \"[specific improvement for top gap]\"" }, { "title": "100-Point Quality Rubric", "body": "DimensionWeight0-2 (Poor)3-5 (Basic)6-8 (Good)9-10 (Excellent)Design15%No blueprint, ad hocBasic flow documentedFull blueprint with error handlingBlueprint + edge cases + optimizationReliability20%No error handlingBasic retriesDLQ + circuit breaker + fallbackSelf-healing + auto-scalingTesting15%No testsHappy path onlyFull test pyramidChaos testing + load testingMonitoring15%No visibilityBasic success/fail logsDashboard + alertsPredictive monitoringDocumentation10%NoneREADME existsBlueprint + runbookFull docs + training materialsSecurity10%Hardcoded credentialsEncrypted secretsLeast privilege + rotationZero-trust + audit trailPerformance10%Works but slowAcceptable speedOptimized + cachedAuto-scaling + sub-secondGovernance5%No registryListed somewhereFull registry + reviewsChange management + compliance\n\nScore: (weighted sum) → Grade: A (90+) B (80-89) C (70-79) D (60-69) F (<60)" }, { "title": "10 Automation Killers", "body": "#MistakeFix1Automating a broken processFix the process FIRST, then automate2No error handlingEvery step needs a failure path3Silent failuresIf it fails and nobody knows, it's worse than manual4Not testing edge casesTest empty, duplicate, malformed, concurrent5Hardcoded valuesUse config/environment variables for everything6No monitoringYou can't fix what you can't see7Building monolith workflowsOne workflow, one job. Chain them together8Ignoring rate limitsDesign for API limits from day one9No documentationFuture-you will hate present-you10Over-automatingNot everything should be automated. Human judgment exists for a reason" }, { "title": "Small Team / Solo Founder", "body": "Start with Zapier/Make — speed over flexibility\nAutomate the 3 most time-consuming tasks first\nGraduate to n8n when spending >$100/mo on no-code" }, { "title": "Regulated Industry", "body": "Add approval gates at every decision point\nLog all automated actions for audit trail\nReview automations quarterly with compliance team\nDocument data flow for privacy impact assessments" }, { "title": "Legacy Systems", "body": "Use middleware/iPaaS for legacy integration\nBuild adapters that normalize legacy data formats\nPlan for eventual migration, not permanent workarounds" }, { "title": "Multi-Team / Enterprise", "body": "Establish automation Center of Excellence (CoE)\nStandardize on 1-2 platforms max\nShared component library for common patterns\nGovernance board for cross-team automations" }, { "title": "AI-Heavy Workflows", "body": "Always keep human-in-the-loop for high-stakes decisions\nMonitor AI output quality continuously\nBudget for AI API costs separately (they scale differently)\nVersion-pin AI models — don't auto-upgrade in production" }, { "title": "Natural Language Commands", "body": "Use these to invoke specific phases:\n\naudit my processes for automation opportunities → Phase 1\nprioritize automations by ROI → Phase 2\nrecommend automation platform for [process] → Phase 3\ndesign workflow blueprint for [process] → Phase 4\nplan integration between [system A] and [system B] → Phase 5\ndesign error handling for [workflow] → Phase 6\ncreate test plan for [automation] → Phase 7\nset up monitoring for [workflow] → Phase 8\noptimize [workflow] for scale → Phase 9\nreview automation governance → Phase 10\nadd AI to [workflow] → Phase 11\nassess automation maturity → Phase 12" } ], "body": "Business Automation Strategy — AfrexAI\n\nThe complete methodology for identifying, designing, building, and scaling business automations. Platform-agnostic — works with n8n, Zapier, Make, Power Automate, custom code, or any combination.\n\nPhase 1: Automation Audit — Find the Gold\n\nBefore building anything, map where time and money leak.\n\nQuick ROI Triage\n\nAsk these 5 questions about any process:\n\nHow often does it happen? (frequency)\nHow long does it take? (duration per occurrence)\nHow many people touch it? (handoffs)\nHow error-prone is it? (failure rate)\nHow much does failure cost? (impact)\nProcess Inventory Template\nprocess_inventory:\n process_name: \"[Name]\"\n department: \"[Sales/Marketing/Ops/Finance/HR/Engineering]\"\n owner: \"[Person responsible]\"\n frequency: \"[X per day/week/month]\"\n duration_minutes: [time per occurrence]\n monthly_volume: [total occurrences]\n monthly_hours: [volume × duration ÷ 60]\n hourly_cost: [fully loaded employee cost]\n monthly_cost: \"$[hours × hourly cost]\"\n error_rate: \"[X%]\"\n error_cost_per_incident: \"$[average]\"\n handoffs: [number of people involved]\n current_tools: [\"tool1\", \"tool2\"]\n automation_potential: \"[Full/Partial/Assist/None]\"\n complexity: \"[Simple/Medium/Complex/Enterprise]\"\n dependencies: [\"system1\", \"system2\"]\n notes: \"[Pain points, workarounds, tribal knowledge]\"\n\nAutomation Potential Classification\nLevel\tDescription\tHuman Role\tExample\nFull\tEnd-to-end automated, no human needed\tMonitor exceptions\tInvoice processing, data sync\nPartial\tAutomated with human approval gates\tReview & approve\tContract generation, hiring workflow\nAssist\tHuman does work, automation helps\tExecute with AI assistance\tCustomer support, content creation\nNone\tRequires human judgment/creativity\tFull ownership\tStrategy, relationship building\nROI Calculation\nAnnual savings = (monthly_hours × 12 × hourly_cost) + (error_rate × volume × 12 × error_cost)\nBuild cost = development_hours × developer_rate + tool_costs\nPayback period = build_cost ÷ (annual_savings ÷ 12) months\nROI = ((annual_savings - annual_tool_cost) ÷ build_cost) × 100%\n\n\nDecision rules:\n\nPayback < 3 months → Build immediately\nPayback 3-6 months → Build this quarter\nPayback 6-12 months → Evaluate against alternatives\nPayback > 12 months → Reconsider (unless strategic)\nPhase 2: Prioritization — The Automation Stack Rank\nICE-R Scoring (0-10 each)\nDimension\tWeight\tScoring Guide\nImpact\t30%\t10=saves >$50K/yr, 7=saves >$20K/yr, 5=saves >$5K/yr, 3=saves >$1K/yr\nConfidence\t20%\t10=proven pattern, 7=similar done before, 5=feasible but new, 3=uncertain\nEase\t25%\t10=<1 day, 7=<1 week, 5=<1 month, 3=<3 months, 1=>3 months\nReliability\t25%\t10=deterministic, 7=95%+ success, 5=80%+ success, 3=needs frequent fixes\nScore = (Impact × 0.30) + (Confidence × 0.20) + (Ease × 0.25) + (Reliability × 0.25)\n\nQuick Win Identification\n\nAutomate FIRST (highest ROI, lowest risk):\n\nData entry / copy-paste between systems\nNotification routing (email → Slack → SMS based on rules)\nReport generation and distribution\nFile organization and naming\nStatus updates across tools\nMeeting scheduling and follow-ups\nInvoice creation from templates\nLead capture → CRM entry\nOnboarding checklists\nBackup and archival\n\nAutomate LAST (complex, high risk):\n\nAnything involving money transfers without approval\nCustomer-facing responses without review\nLegal/compliance decisions\nHiring/firing workflows\nSecurity-sensitive operations\nPhase 3: Platform Selection — Choose Your Weapons\nPlatform Decision Matrix\nFactor\tNo-Code (Zapier/Make)\tLow-Code (n8n/Power Automate)\tCustom Code\tAI Agent\nBest for\tSimple integrations\tComplex workflows\tUnique logic\tJudgment calls\nBuild speed\tHours\tDays\tWeeks\tDays-weeks\nMaintenance\tLow\tMedium\tHigh\tMedium\nFlexibility\tLimited\tHigh\tUnlimited\tHigh\nCost at scale\tExpensive\tModerate\tCheap\tVaries\nError handling\tBasic\tGood\tFull control\tVariable\nTeam skill needed\tBusiness user\tTechnical BA\tDeveloper\tAI engineer\nVendor lock-in\tHigh\tMedium\tNone\tLow-medium\nSelection Decision Tree\nIs the process deterministic (same input → same output)?\n├── YES: Does it involve >3 systems?\n│ ├── YES: Does it need complex branching logic?\n│ │ ├── YES → Low-code (n8n/Power Automate)\n│ │ └── NO → No-code (Zapier/Make) if budget allows, else n8n\n│ └── NO: Is it performance-critical?\n│ ├── YES → Custom code\n│ └── NO → No-code (simplest wins)\n└── NO: Does it need judgment/reasoning?\n ├── YES: Is the judgment pattern learnable?\n │ ├── YES → AI agent with human review\n │ └── NO → Human-assisted automation\n └── NO → Partial automation with human gates\n\nCost Comparison by Scale\nMonthly Tasks\tZapier\tMake\tn8n (self-hosted)\tCustom Code\n1,000\t$30\t$10\t$5 (hosting)\t$50+ (hosting)\n10,000\t$100\t$30\t$5\t$50+\n100,000\t$500+\t$150\t$10\t$50+\n1,000,000\t$2,000+\t$500+\t$20\t$100+\n\nRule: If you're spending >$200/mo on Zapier/Make, evaluate self-hosted n8n.\n\nPhase 4: Workflow Architecture — Design Before You Build\nWorkflow Blueprint Template\nworkflow_blueprint:\n name: \"[Descriptive name]\"\n id: \"WF-[DEPT]-[NUMBER]\"\n version: \"1.0.0\"\n owner: \"[Person]\"\n priority: \"[P0-P3]\"\n \n trigger:\n type: \"[webhook/schedule/event/manual/condition]\"\n source: \"[System or schedule]\"\n conditions: \"[When to fire]\"\n dedup_strategy: \"[How to prevent double-processing]\"\n \n inputs:\n - name: \"[field]\"\n type: \"[string/number/date/object]\"\n required: true\n validation: \"[rules]\"\n source: \"[where it comes from]\"\n \n steps:\n - id: \"step_1\"\n action: \"[verb: fetch/transform/validate/send/create/update/delete]\"\n system: \"[target system]\"\n description: \"[what this step does]\"\n input: \"[from trigger or previous step]\"\n output: \"[what it produces]\"\n error_handling: \"[retry/skip/alert/abort]\"\n timeout_seconds: 30\n \n - id: \"step_2_branch\"\n type: \"condition\"\n condition: \"[expression]\"\n true_path: \"step_3a\"\n false_path: \"step_3b\"\n \n error_handling:\n retry_policy:\n max_attempts: 3\n backoff: \"exponential\"\n initial_delay_seconds: 5\n on_failure: \"[alert/queue-for-review/fallback]\"\n alert_channel: \"[Slack/email/SMS]\"\n dead_letter_queue: true\n \n monitoring:\n success_metric: \"[what defines success]\"\n expected_duration_seconds: [max]\n alert_on_duration_exceeded: true\n log_level: \"[info/debug/error]\"\n \n testing:\n test_data: \"[how to generate test inputs]\"\n expected_output: \"[what success looks like]\"\n edge_cases: [\"empty input\", \"duplicate\", \"malformed data\"]\n\n7 Workflow Design Principles\nIdempotent by default — Running the same workflow twice with the same input should produce the same result, not duplicates\nFail loudly — Silent failures are worse than crashes. Every error must notify someone\nCheckpoint progress — Long workflows should save state so they can resume, not restart\nValidate early — Check inputs at the start, not after 10 expensive API calls\nSeparate concerns — One workflow, one job. Chain workflows, don't build monoliths\nLog everything — Timestamps, inputs, outputs, decisions. You WILL need to debug\nHuman escape hatch — Every automated workflow needs a manual override path\nCommon Workflow Patterns\nPattern\tWhen to Use\tExample\nSequential\tSteps depend on each other\tLead → Enrich → Score → Route\nParallel fan-out\tIndependent steps\tSend email + Update CRM + Log analytics\nConditional branch\tDifferent paths by data\tHigh value → Sales, Low value → Nurture\nLoop/batch\tProcess collections\tFor each row in CSV, create record\nApproval gate\tHuman judgment needed\tContract review before sending\nEvent-driven chain\tWorkflow triggers workflow\tOrder placed → Fulfillment → Shipping → Notification\nRetry with fallback\tUnreliable external APIs\tTry API → Retry 3x → Use cached data → Alert\nScheduled sweep\tPeriodic cleanup/sync\tNightly: sync CRM → accounting\nPhase 5: Integration Architecture — Connect Everything\nIntegration Quality Checklist\n\nFor every system integration:\n\n API documentation reviewed\n Authentication method confirmed (OAuth2/API key/JWT)\n Rate limits documented (requests/min, requests/day)\n Webhook support checked (push vs poll)\n Error response format understood\n Pagination handling planned\n Data format confirmed (JSON/XML/CSV)\n Field mapping documented\n Test environment available\n Sandbox/production separation configured\nData Mapping Template\ndata_mapping:\n source_system: \"[System A]\"\n target_system: \"[System B]\"\n sync_direction: \"[one-way/bidirectional]\"\n sync_frequency: \"[real-time/5min/hourly/daily]\"\n conflict_resolution: \"[source wins/target wins/newest wins/manual]\"\n \n field_mappings:\n - source_field: \"contact.email\"\n target_field: \"customer.email_address\"\n transform: \"lowercase\"\n required: true\n - source_field: \"contact.company\"\n target_field: \"customer.organization\"\n transform: \"trim\"\n default: \"Unknown\"\n - source_field: \"contact.created_at\"\n target_field: \"customer.signup_date\"\n transform: \"ISO8601 → YYYY-MM-DD\"\n\nRate Limit Strategy\nApproach\tWhen\tImplementation\nQueue + throttle\tPredictable volume\tProcess queue at 80% of rate limit\nExponential backoff\tBurst traffic\tWait 1s, 2s, 4s, 8s on 429 errors\nBatch API calls\tHigh volume CRUD\tGroup 50-100 records per call\nCache responses\tRepeated lookups\tCache for TTL matching data freshness needs\nOff-peak scheduling\tNon-urgent syncs\tRun heavy syncs at 2-4 AM\nPhase 6: Error Handling & Reliability — Build It Unbreakable\nError Classification\nType\tExample\tResponse\tPriority\nTransient\tAPI timeout, 503\tRetry with backoff\tAuto-handle\nRate limit\t429 Too Many Requests\tQueue + throttle\tAuto-handle\nData validation\tMissing required field\tLog + skip + alert\tReview daily\nAuth failure\tToken expired\tRefresh + retry, else alert\tP1 — fix within 1h\nLogic error\tUnexpected state\tHalt + alert + queue\tP0 — fix immediately\nExternal change\tAPI schema changed\tHalt + alert\tP0 — fix immediately\nCapacity\tQueue overflow\tScale + alert\tP1 — fix within 4h\nDead Letter Queue Pattern\n\nEvery workflow should have a DLQ:\n\nCapture — Failed items go to DLQ with full context (input, error, timestamp, step)\nAlert — Notify on DLQ growth (>10 items or >1% failure rate)\nReview — Daily check of DLQ items\nReplay — Ability to reprocess DLQ items after fix\nExpire — Auto-archive items older than 30 days with summary\nCircuit Breaker Pattern\nStates: CLOSED (normal) → OPEN (failing) → HALF-OPEN (testing)\n\nCLOSED: Process normally, track failures\n → If failure_count > threshold in window → OPEN\n\nOPEN: Reject all requests, return cached/default\n → After cool_down_period → HALF-OPEN\n\nHALF-OPEN: Allow 1 test request\n → If success → CLOSED\n → If failure → OPEN (reset cool_down)\n\n\nThresholds:\n\nSimple integrations: 5 failures in 60 seconds\nCritical paths: 3 failures in 30 seconds\nNon-critical: 10 failures in 300 seconds\nPhase 7: Testing & Validation — Trust But Verify\nAutomation Test Pyramid\nLevel\tWhat\tHow\tWhen\nUnit\tIndividual step logic\tMock inputs, verify output\tEvery change\nIntegration\tSystem connections\tTest with sandbox APIs\tWeekly + after changes\nEnd-to-end\tFull workflow path\tRun with test data\tBefore deploy + weekly\nChaos\tFailure scenarios\tKill steps, corrupt data\tMonthly\nLoad\tVolume handling\t10x normal volume\tBefore scaling\nTest Scenario Checklist\n\nFor every workflow, test:\n\n Happy path (normal input, expected output)\n Empty/null input (missing required fields)\n Duplicate input (same event twice)\n Malformed input (wrong types, encoding issues)\n Boundary values (max length, zero, negative)\n API down (target system unavailable)\n Slow response (timeout handling)\n Partial failure (step 3 of 5 fails)\n Concurrent execution (two runs at same time)\n Clock skew / timezone issues\n Large payload (oversized data)\n Permission denied (auth issues)\nValidation Before Go-Live\ngo_live_checklist:\n functionality:\n - [ ] All test scenarios pass\n - [ ] Edge cases documented and handled\n - [ ] Error messages are actionable\n \n reliability:\n - [ ] Retry logic tested\n - [ ] Circuit breaker configured\n - [ ] Dead letter queue active\n - [ ] Idempotency verified (run twice, same result)\n \n monitoring:\n - [ ] Success/failure alerts configured\n - [ ] Duration alerts set\n - [ ] Log retention configured\n - [ ] Dashboard created\n \n documentation:\n - [ ] Workflow blueprint updated\n - [ ] Runbook written\n - [ ] Team trained on manual override\n \n rollback:\n - [ ] Previous version preserved\n - [ ] Rollback procedure tested\n - [ ] Data cleanup plan for partial runs\n\nPhase 8: Monitoring & Observability — See Everything\nAutomation Health Dashboard\nautomation_dashboard:\n period: \"weekly\"\n \n summary:\n total_workflows: [count]\n total_executions: [count]\n success_rate: \"[X%]\"\n avg_duration: \"[X seconds]\"\n errors_this_period: [count]\n time_saved_hours: [calculated]\n cost_saved: \"$[calculated]\"\n \n by_workflow:\n - name: \"[Workflow name]\"\n executions: [count]\n success_rate: \"[X%]\"\n avg_duration: \"[X seconds]\"\n p95_duration: \"[X seconds]\"\n errors: [count]\n error_types: [\"type1: count\", \"type2: count\"]\n dlq_items: [count]\n status: \"[healthy/degraded/failing]\"\n \n alerts_fired: [count]\n manual_interventions: [count]\n \n top_issues:\n - \"[Issue 1: description + fix status]\"\n - \"[Issue 2: description + fix status]\"\n \n cost:\n platform_cost: \"$[monthly]\"\n api_calls_cost: \"$[monthly]\"\n compute_cost: \"$[monthly]\"\n total: \"$[monthly]\"\n cost_per_execution: \"$[calculated]\"\n\nAlert Rules\nMetric\tWarning\tCritical\tAction\nSuccess rate\t<95%\t<90%\tInvestigate + fix\nDuration\t>2x average\t>5x average\tCheck for bottleneck\nDLQ size\t>10 items\t>50 items\tReview + reprocess\nError spike\t5 errors/hour\t20 errors/hour\tPause + investigate\nQueue depth\t>100 pending\t>1000 pending\tScale + investigate\nCost spike\t>150% of average\t>300% of average\tAudit + optimize\nWeekly Review Questions\nWhich workflows had the lowest success rate? Why?\nAre any workflows consistently slow? What's the bottleneck?\nHow many manual interventions were needed? Can we eliminate them?\nWhat's in the DLQ? Patterns?\nAre we approaching any rate limits?\nTotal cost vs total time saved — still positive ROI?\nPhase 9: Scaling & Optimization — Go From 10 to 10,000\nScaling Checklist\n\nBefore scaling any automation:\n\n Load tested at 10x current volume\n Rate limits mapped for all APIs\n Queue-based architecture (not synchronous chains)\n Database indexes optimized\n Caching layer in place\n Monitoring alerts adjusted for new thresholds\n Cost projections at scale calculated\n Fallback/degradation plan documented\nPerformance Optimization Priority\nEliminate unnecessary API calls — Cache lookups, batch operations\nParallelize independent steps — Don't wait when you don't have to\nOptimize data payloads — Only fetch/send fields you need\nUse webhooks over polling — Real-time + fewer API calls\nBatch processing — Group operations (50-100 per batch)\nAsync where possible — Don't block on non-critical steps\nCDN/cache for static lookups — Country codes, categories, templates\nDatabase query optimization — Indexes, query plans, connection pooling\nWhen to Migrate Platforms\nSignal\tFrom\tTo\nSpending >$500/mo on Zapier/Make\tNo-code\tSelf-hosted n8n\nNeed custom logic in >50% of workflows\tNo-code\tLow-code or code\n>100K executions/day\tAny hosted\tSelf-hosted or custom\nComplex branching breaking visual tools\tLow-code\tCustom code\nMultiple teams building automations\tSingle tool\tPlatform + governance\nAI judgment needed in workflows\tTraditional\tAI agent integration\nPhase 10: Governance & Documentation — Keep It Manageable\nAutomation Registry\n\nEvery automation must be registered:\n\nautomation_registry_entry:\n id: \"WF-[DEPT]-[NUMBER]\"\n name: \"[Descriptive name]\"\n description: \"[What it does in one sentence]\"\n owner: \"[Person]\"\n team: \"[Department]\"\n platform: \"[n8n/Zapier/Make/custom]\"\n status: \"[active/paused/deprecated/testing]\"\n created: \"[date]\"\n last_modified: \"[date]\"\n last_reviewed: \"[date]\"\n review_frequency: \"[monthly/quarterly]\"\n \n business_impact:\n time_saved_monthly_hours: [X]\n cost_saved_monthly: \"$[X]\"\n error_reduction: \"[X%]\"\n \n technical:\n trigger: \"[type]\"\n systems_connected: [\"system1\", \"system2\"]\n avg_daily_executions: [X]\n success_rate: \"[X%]\"\n \n dependencies:\n upstream: [\"WF-XXX\"]\n downstream: [\"WF-YYY\"]\n \n documentation:\n blueprint: \"[link]\"\n runbook: \"[link]\"\n test_plan: \"[link]\"\n\nNaming Conventions\nPattern: [DEPT]-[ACTION]-[OBJECT]-[QUALIFIER]\nExamples:\n SALES-sync-leads-from-typeform\n FINANCE-generate-invoice-monthly\n HR-onboard-employee-new-hire\n MARKETING-post-content-social-scheduled\n OPS-backup-database-nightly\n\nChange Management for Automations\nChange Type\tApproval\tTesting\tRollback Plan\nConfig change (threshold, timing)\tOwner\tQuick smoke test\tRevert config\nLogic change (new branch, new step)\tOwner + reviewer\tFull test suite\tPrevious version\nIntegration change (new API, new system)\tOwner + tech lead\tIntegration + E2E\tDisconnect + manual\nNew workflow\tOwner + stakeholder\tFull test + pilot\tDisable workflow\nDeprecation\tOwner + affected teams\tVerify replacements\tRe-enable\nQuarterly Automation Review\nInventory check — Are all automations in the registry? Any rogue workflows?\nROI validation — Is each automation still delivering value?\nHealth review — Success rates, error trends, DLQ patterns\nCost audit — Platform costs trending up? Optimization opportunities?\nSecurity review — API keys rotated? Permissions still appropriate?\nDeprecation candidates — Any automations that should be retired?\nOpportunity scan — New processes to automate? Existing ones to improve?\nPhase 11: AI-Powered Automations — The Next Level\nWhen to Add AI to Automations\nScenario\tAI Type\tExample\nClassify unstructured text\tLLM\tCategorize support tickets\nExtract data from documents\tLLM + OCR\tParse invoices, contracts\nGenerate content from templates\tLLM\tPersonalized emails, reports\nMake judgment calls\tLLM + rules\tLead scoring, risk assessment\nSummarize information\tLLM\tMeeting notes, research briefs\nRoute based on intent\tLLM\tCustomer request → right team\nAI Integration Best Practices\nAlways validate AI output — LLMs hallucinate. Add validation checks\nSet confidence thresholds — Below threshold → human review queue\nLog AI decisions — Input, output, confidence, model version\nA/B test AI vs rules — Prove AI adds value before committing\nCost-control AI calls — Cache similar inputs, batch where possible\nFallback to rules — If AI is unavailable, have deterministic backup\nReview AI decisions weekly — Spot check for quality drift\nAI Agent Integration Pattern\nai_agent_step:\n type: \"ai_judgment\"\n model: \"[model name]\"\n \n input:\n context: \"[relevant data from previous steps]\"\n task: \"[specific instruction — be precise]\"\n output_format: \"[JSON schema or structured format]\"\n constraints: [\"must not\", \"must always\", \"if unsure\"]\n \n validation:\n confidence_threshold: 0.85\n required_fields: [\"field1\", \"field2\"]\n value_ranges:\n score: [0, 100]\n category: [\"A\", \"B\", \"C\"]\n \n on_low_confidence:\n action: \"route_to_human\"\n queue: \"[review queue name]\"\n \n on_failure:\n action: \"fallback_to_rules\"\n rules_engine: \"[rule set name]\"\n \n monitoring:\n log_all_decisions: true\n sample_rate_for_review: 0.10\n alert_on_confidence_drop: true\n\nPhase 12: Automation Maturity Model\n5 Levels of Automation Maturity\nLevel\tName\tDescription\tIndicators\n1\tAd Hoc\tManual processes, maybe a few scripts\tNo registry, tribal knowledge\n2\tReactive\tAutomate pain points as they arise\tSome workflows, no standards\n3\tSystematic\tPlanned automation program\tRegistry, testing, monitoring\n4\tOptimized\tContinuous improvement, governance\tROI tracking, quarterly reviews\n5\tIntelligent\tAI-augmented, self-healing\tAdaptive workflows, predictive\nMaturity Assessment (Score 1-5 per dimension)\nautomation_maturity:\n dimensions:\n strategy: [1-5] # Planned roadmap vs ad hoc\n architecture: [1-5] # Patterns, standards, reuse\n reliability: [1-5] # Error handling, monitoring, uptime\n governance: [1-5] # Registry, change management, reviews\n testing: [1-5] # Test coverage, validation, chaos\n documentation: [1-5] # Blueprints, runbooks, training\n optimization: [1-5] # Performance, cost, continuous improvement\n ai_integration: [1-5] # AI-powered decisions, self-healing\n \n total: [sum ÷ 8]\n grade: \"[A/B/C/D/F]\"\n # A: 4.5+ | B: 3.5-4.4 | C: 2.5-3.4 | D: 1.5-2.4 | F: <1.5\n \n top_gap: \"[lowest scoring dimension]\"\n next_action: \"[specific improvement for top gap]\"\n\n100-Point Quality Rubric\nDimension\tWeight\t0-2 (Poor)\t3-5 (Basic)\t6-8 (Good)\t9-10 (Excellent)\nDesign\t15%\tNo blueprint, ad hoc\tBasic flow documented\tFull blueprint with error handling\tBlueprint + edge cases + optimization\nReliability\t20%\tNo error handling\tBasic retries\tDLQ + circuit breaker + fallback\tSelf-healing + auto-scaling\nTesting\t15%\tNo tests\tHappy path only\tFull test pyramid\tChaos testing + load testing\nMonitoring\t15%\tNo visibility\tBasic success/fail logs\tDashboard + alerts\tPredictive monitoring\nDocumentation\t10%\tNone\tREADME exists\tBlueprint + runbook\tFull docs + training materials\nSecurity\t10%\tHardcoded credentials\tEncrypted secrets\tLeast privilege + rotation\tZero-trust + audit trail\nPerformance\t10%\tWorks but slow\tAcceptable speed\tOptimized + cached\tAuto-scaling + sub-second\nGovernance\t5%\tNo registry\tListed somewhere\tFull registry + reviews\tChange management + compliance\n\nScore: (weighted sum) → Grade: A (90+) B (80-89) C (70-79) D (60-69) F (<60)\n\n10 Automation Killers\n#\tMistake\tFix\n1\tAutomating a broken process\tFix the process FIRST, then automate\n2\tNo error handling\tEvery step needs a failure path\n3\tSilent failures\tIf it fails and nobody knows, it's worse than manual\n4\tNot testing edge cases\tTest empty, duplicate, malformed, concurrent\n5\tHardcoded values\tUse config/environment variables for everything\n6\tNo monitoring\tYou can't fix what you can't see\n7\tBuilding monolith workflows\tOne workflow, one job. Chain them together\n8\tIgnoring rate limits\tDesign for API limits from day one\n9\tNo documentation\tFuture-you will hate present-you\n10\tOver-automating\tNot everything should be automated. Human judgment exists for a reason\nEdge Cases\nSmall Team / Solo Founder\nStart with Zapier/Make — speed over flexibility\nAutomate the 3 most time-consuming tasks first\nGraduate to n8n when spending >$100/mo on no-code\nRegulated Industry\nAdd approval gates at every decision point\nLog all automated actions for audit trail\nReview automations quarterly with compliance team\nDocument data flow for privacy impact assessments\nLegacy Systems\nUse middleware/iPaaS for legacy integration\nBuild adapters that normalize legacy data formats\nPlan for eventual migration, not permanent workarounds\nMulti-Team / Enterprise\nEstablish automation Center of Excellence (CoE)\nStandardize on 1-2 platforms max\nShared component library for common patterns\nGovernance board for cross-team automations\nAI-Heavy Workflows\nAlways keep human-in-the-loop for high-stakes decisions\nMonitor AI output quality continuously\nBudget for AI API costs separately (they scale differently)\nVersion-pin AI models — don't auto-upgrade in production\nNatural Language Commands\n\nUse these to invoke specific phases:\n\naudit my processes for automation opportunities → Phase 1\nprioritize automations by ROI → Phase 2\nrecommend automation platform for [process] → Phase 3\ndesign workflow blueprint for [process] → Phase 4\nplan integration between [system A] and [system B] → Phase 5\ndesign error handling for [workflow] → Phase 6\ncreate test plan for [automation] → Phase 7\nset up monitoring for [workflow] → Phase 8\noptimize [workflow] for scale → Phase 9\nreview automation governance → Phase 10\nadd AI to [workflow] → Phase 11\nassess automation maturity → Phase 12" }, "trust": { "sourceLabel": "tencent", "provenanceUrl": "https://clawhub.ai/1kalin/afrexai-automation-strategy", "publisherUrl": "https://clawhub.ai/1kalin/afrexai-automation-strategy", "owner": "1kalin", "version": "1.0.0", "license": null, "verificationStatus": "Indexed source record" }, "links": { "detailUrl": "https://openagent3.xyz/skills/afrexai-automation-strategy", "downloadUrl": "https://openagent3.xyz/downloads/afrexai-automation-strategy", "agentUrl": "https://openagent3.xyz/skills/afrexai-automation-strategy/agent", "manifestUrl": "https://openagent3.xyz/skills/afrexai-automation-strategy/agent.json", "briefUrl": "https://openagent3.xyz/skills/afrexai-automation-strategy/agent.md" } }