The five-stage pipeline
Stage-by-stage breakdown
1 · Lead Intake
What happens: AgentFlow captures inbound lead data through a configurable intake form or a signed inbound webhook. Every submission is timestamped and stored behind authenticated access.Who is involved: The lead (submitter) and the platform. No operator action is required at this stage.Output: A structured lead record with all submitted fields, a unique ID, and a
pending_qualification status that triggers the next stage automatically.2 · AI Qualification
What happens: OpenAI processes the lead record against your configured qualification criteria. The engine returns a confidence score (0–100), a fit label (strong / moderate / weak), and a plain-language qualification summary explaining the reasoning.Who is involved: The AI engine and your qualification schema. No operator action is required until the output is ready.Output: An enriched lead record with the AI’s structured qualification result attached, surfaced in the operator dashboard for review.
3 · Operator Review
What happens: The qualified lead enters the operator queue. The reviewing operator sees the original lead data alongside the AI output and makes one of three decisions: approve for billing, flag for follow-up, or reject. Every decision is logged with a timestamp and operator ID.Who is involved: Your designated operator or review team.Output: A reviewed lead record with an operator decision, a decision timestamp, and an audit log entry—ready for the billing stage.
4 · Billing Readiness
What happens: Approved leads trigger the billing readiness check. If Stripe or PayPal is configured, this stage surfaces the relevant checkout link, invoice action, or payment intent for the operator to dispatch. The stage is EU VAT-aware and operates in either test or live mode depending on your provider configuration.Who is involved: The operator, the billing provider, and (optionally) the lead.Output: A billing-ready lead record with a payment action or a confirmation that no billing step is required before handoff.
5 · CRM / Workflow Handoff
What happens: The qualified, reviewed lead is packaged and dispatched to your configured downstream destination—a CRM record, a Slack notification, an outbound webhook, or a spreadsheet row. The handoff is operator-controlled: you define the trigger condition and destination in the Integrations panel.Who is involved: The platform’s integration layer and your downstream system.Output: A completed pipeline record marked
handed_off, with delivery confirmation and a reference ID from the downstream system.Cross-stage: Audit & Observability
What happens: Every stage transition is logged—submission timestamp, AI call result, operator decision, billing action, and handoff confirmation. The audit trail is available in the operator dashboard and can be exported for compliance or client reporting purposes.Who is involved: The platform automatically; operators can query or export logs at any time.Output: A complete, timestamped event log for every lead that has passed through the pipeline.
Key integrations
| Integration | Role in the pipeline | Stage |
|---|---|---|
| OpenAI | Powers lead scoring, fit labeling, and qualification summaries | AI Qualification |
| Stripe | Provides checkout readiness, payment intents, and invoice actions | Billing Readiness |
| PayPal | Alternative payment path for checkout and invoice dispatch | Billing Readiness |
| CRM / Webhook | Receives the qualified lead at handoff via outbound webhook or direct integration | CRM Handoff |
All integrations require you to configure your own provider accounts with live credentials. The platform’s billing and payment stages operate in test mode until you verify your Stripe or PayPal live-mode configuration. Review your provider documentation alongside AgentFlow’s integration guides before accepting real payments.
Who benefits
AgentFlow is designed for operators who need a structured path from first lead signal to revenue action—not another tool that adds steps without removing uncertainty.Agencies
Package AgentFlow as an AI RevOps foundation for client delivery. The five-stage pipeline gives client teams a repeatable qualification workflow without building it from scratch. Each deployment can be scoped to the client’s CRM, payment provider, and qualification criteria.
RevOps Consultants
Structure repeatable client qualification processes. AgentFlow sits between lead capture and operational follow-through, giving you a configurable layer that enforces consistent scoring, review, and handoff standards across every engagement.
SaaS Founders
Validate a productized qualification workflow before investing in a custom build. Use the foundation to run your first qualified demo requests, paid discovery calls, or onboarding screens through a structured pipeline rather than manual triage.
Senior Developers
Evaluate a production-conscious SaaS foundation built on OpenAI, Stripe/PayPal, and modern cloud infrastructure. The platform’s webhook safety model, authentication boundaries, and operator-controlled pipeline are designed to meet the standards technical teams expect before committing to a build.
Architecture principles
AgentFlow is built around three design principles that apply across every stage of the pipeline:Operator control at every decision point
Operator control at every decision point
AI outputs are inputs to human decisions, not replacements for them. The qualification engine produces a score and a summary; the operator approves, flags, or rejects. No lead advances to billing or handoff without an explicit operator action. This keeps your team accountable and your audit trail clean.
Trust boundaries on all external events
Trust boundaries on all external events
Inbound webhooks are treated as untrusted until signature-verified. Provider events (payment confirmations, form submissions from third-party tools) are validated server-side before they touch the pipeline. This principle applies to every integration point, not just payment webhooks.
Configuration over assumption
Configuration over assumption
AgentFlow ships with a pipeline shape, not with your business rules baked in. You define qualification criteria, approval thresholds, billing triggers, and handoff destinations through configuration—so the system matches your workflow rather than forcing you to match its defaults.
Next steps
Dive deeper into the concepts and guides that map directly to each pipeline stage:Lead Qualification
Understand how AgentFlow structures qualification criteria, what fields the AI engine expects, and how to tune scoring for your lead type.
Workflow Pipeline
Explore the full pipeline architecture, stage-transition rules, and how status changes are tracked across the data layer.
Operator Review
Learn how the operator queue works, how to configure review assignments, and what the audit log captures at each decision point.
Billing Readiness
Configure Stripe or PayPal, move from test to live mode, and understand EU VAT handling before you process your first payment.