Design stages and ownership
A stage should describe the Lead’s current business condition and the next job to be done. It should not merely name a channel, employee, or completed task.
Required structure
Every pipeline needs:
- one Entry stage, kept first;
- one or more Working or Closing stages for active work;
- one Won stage with 100% probability;
- one Lost stage with 0% probability.
FitStack places Won and Lost at the end when the pipeline is saved.
Disqualified is also a terminal Lead status, but it is not a pipeline stage. Staff can use it to close an opportunity that should not proceed without moving the Lead to Won or Lost.
| Stage type | Use it for | Example |
|---|---|---|
| Entry | the first accepted state of a new opportunity | New Lead |
| Working | discovery, contact, nurture, or qualification | AI Engagement |
| Closing | a late-stage commitment or decision | Tour Booked or Offer Presented |
| Won | the intended business outcome | Membership Started |
| Lost | a closed outcome that did not convert | Not Interested |
Give every stage a contract
For each nonterminal stage, write down:
- entry condition — why a Lead belongs here;
- next job — what must happen now;
- driver — human staff or one configured AI agent;
- evidence of success — the observable condition that justifies movement;
- success destination — the next stage;
- failure or no-response plan — another stage, manual review, or stale policy;
- time expectation — how long the Lead may remain before becoming stale.
Use the probability as a consistent forecast estimate for Leads in that stage. It does not move Leads or prove conversion by itself.
Choose Manual or agentic ownership
Use Manual when the work requires direct human judgment, a sensitive conversation, an unsupported channel, or no repeatable cadence.
Assign an AI workflow when the stage has a repeatable goal, approved channels, clear success evidence, and an explicit human exception path. One stage can have only one active workflow and one selected agent.
Design stale protections
The pipeline has a default number of days before a Lead becomes stale. A stage can override that value. Then choose what happens after staleness:
- Do nothing — mark it stale for human review;
- Move to stage — route it to a recovery or review stage;
- Mark as Lost — close it after the optional delay.
Stale timing and workflow timing are independent. Make the stale window longer than the expected workflow duration. Otherwise FitStack can mark or move the Lead as stale while the configured AI workflow is still running.
Review paths before saving
Trace at least these scenarios from Entry to a terminal or owned state:
- the person responds positively;
- the person responds but the goal is not reached;
- there is no response;
- a staff member intervenes;
- the Lead becomes stale;
- the Lead should be Won;
- the Lead should be Lost;
- the Lead should be Disqualified.
FitStack warns about cycles, but the warning does not replace this review. A loop can be intentional only when timing, ownership, and an eventual exit are clear.