What is the difference between a workflow rule and a flow in Strkr?
A workflow rule is simple if-then automation that fires on a trigger (a field changes, a record is created, a timer elapses) and runs a specific action (update a field, send an email, create a task). A flow is a multi-step orchestration with specific primitives: branching, typed variables, external API calls, approval gates, AI drafting steps, scheduled actions, error handling, and audit logging. Workflow rules handle the simple end of automation. Flows handle the complex end. Strkr supports both: simple workflow rules for the trivial automation needs ("when a deal closes, send a thank-you note"), and the flow builder for the serious orchestration needs ("when a lead comes in, enrich via external APIs with retry handling, compute qualification based on typed scoring variables, branch based on score plus ICP match, route to the appropriate AE with capacity awareness, create opportunity with specific stage assignment, notify across multiple channels, schedule a follow-up check in 72 hours"). The division of labor keeps simple things simple and makes complex things possible.
Who in the business can configure flows in Strkr?
The flow builder is designed for RevOps leads, operations analysts, specific-admin users, and technically-inclined business users to configure flows without engineering involvement. The visual canvas eliminates the specific requirement for code (versus Salesforce Apex). Typed variables and visual branching logic eliminate the specific requirement for formal programming training. Standard business users with specific operations backgrounds typically become productive on the flow builder within 2 to 4 hours of specific guided onboarding. For specific complex flows (advanced API integrations with specific authentication patterns, specific-complex typed-variable manipulations, specific advanced error handling), engineering involvement may still be helpful but is not required for the vast majority of flow configurations. Role-based permissions control who can edit which flows: the marketing operations user controls marketing flows, the sales operations user controls sales flows, the compliance officer controls compliance-review flows. Changes to production flows can require approval from a designated reviewer per role, which gives the specific operational discipline without blocking agility.
How does Strkr handle flows that need to call external systems with authentication?
The external-API step supports all common authentication patterns: bearer tokens with specific-token-source configuration (static token, OAuth 2.0 token refresh, dynamic token from a prior step), API keys with secret management (the keys store in Strkr's secret vault with specific-rotation support), OAuth 2.0 with full authorization code and client credentials flows, basic authentication, and custom header injection for specific-vendor authentication schemes. Request configuration includes URL with variable substitution, HTTP method, headers with variable substitution, body (JSON, form-encoded, multipart) with variable substitution, timeout configuration, and specific retry logic with exponential backoff. Response handling includes typed JSON parsing with schema validation (so the response structure validates at step time rather than runtime), error-case routing based on status codes (specific routing for 404, 401, 429, 5xx), rate-limit respecting based on vendor-specific headers (Retry-After, X-RateLimit headers), and specific audit logging of the request-response pair with appropriate redaction of sensitive content. The specific benefit versus Zapier or similar integration tools: the external-API logic lives inside the business-logic flow with the full context available, rather than as a separate integration step that handoffs data between tools. For teams with specific requirements (mutual TLS, specific OAuth variations, specific-vendor custom auth), the flow builder supports these through specific configuration without requiring code.
What happens when a flow fails in production?
Flow failures are first-class citizens with specific operational handling. Transient failures (external API timeout, specific-vendor temporary unavailability, specific-network blip) retry automatically with exponential backoff for the configured retry count. If retries exhaust, the specific failure handling depends on the flow configuration. Three common patterns: (1) route to a specific-alternative path (the flow continues through a graceful-degradation branch that handles the failure without stopping the broader motion), (2) escalate to a human for intervention (the flow pauses and routes a specific alert to the operations owner with full context about what failed and why), (3) write to a dead-letter queue (the flow instance halts with full state preservation; the operations owner reviews DLQ entries, resolves the underlying issue, and either reprocesses the specific instance or marks it as handled). All three patterns support specific audit logging so the historical record of what happened is retrievable. For specific compliance-sensitive flows (financial transactions, specific regulatory communications, specific legally-required actions), the failure handling can require specific documented resolution with the specific approver attestation retained. The specific operational benefit is that production failures become managed events rather than silent data corruption that surfaces weeks or months later. Alerting on anomaly patterns (failure rate climbing above baseline, specific-flow success rate dropping, specific-vendor API consistently failing) surfaces systemic issues before they produce broad impact.
Can Strkr's flow builder replace Zapier for serious integration workloads?
For integration workloads that involve the CRM as a primary party (lead flows, deal progression, customer-success automation, specific business-logic flows that live around the CRM data), yes. The flow builder covers the specific capability depth that Zapier provides plus the engineering-grade capabilities (typed variables, approval gates, specific-error handling, idempotency, observability) that Zapier does not provide cleanly. For integration workloads that do not involve the CRM as a primary party (specific-tool-to-tool automation that happens outside the CRM, specific personal-productivity automation that lives in individual-user tools), Zapier or similar lightweight tools may still be the appropriate choice because the CRM is not the right place for that logic. The specific migration pattern that teams follow: identify the specific Zapier zaps that touch the CRM, migrate those to Strkr flows (typically a 2 to 6 week project for teams with 10 to 50 zaps), retire the Zapier subscription if no non-CRM zaps remain, or scale down the Zapier subscription significantly if some zaps remain. The specific economic case plus the operational case (one place for business-critical automation, visibility into flow health, specific-engineering-grade reliability) typically clears the migration project for teams that have grown beyond trivial Zapier usage. We are not suggesting Zapier as the solution for CRM-centric automation; the specific case is that native flows are the right tool for CRM automation while lightweight integration tools may still have a role in non-CRM contexts.
How does Strkr's automation handle AI inside flows?
AI drafting and AI reasoning are first-class step types in the flow builder. The specific capabilities include email drafting (the flow generates an email with the specific-prospect context available), SMS message drafting, call-prep summary generation, specific-content generation (specific-template content with variable context), classification and extraction (parse specific-unstructured input into typed variables for downstream steps), and specific reasoning tasks (summarize this long context, identify the key points, suggest the next action). Model-tier routing applies: simple drafts route to Haiku-class for cost efficiency, complex reasoning routes to Sonnet-class, specific heavy-reasoning tasks route to Opus-class. The AI step reads the live record context, respects the tenant's claims-based authorization (the AI cannot see data the step's executor cannot see), and writes outputs as typed variables for downstream flow consumption. For specific compliance-sensitive flows, the AI output can route through a review step before firing (the AI drafts the communication, the human approver reviews and confirms before send). The AI inference is included in the paid tier without a per-call credit meter below the generous included cap. The specific operational benefit versus external-AI integration via API: the AI step has the full flow context automatically, respects the tenant's permission model natively, and runs with the specific latency and reliability of first-class capability rather than a brittle external-API integration.
