# Pipeline stages, lead scoring and workflows in Linkubit

> How a conversation becomes a scored lead and a priced deal: the starter stages, scoring rules and workflows, and where to change each.

Source: https://www.linkubit.com/docs/pipeline-scoring-and-workflows  
Section: Linkubit help centre › Leads, pipeline and automations  
Updated: 2026-10-17

## In short

Every workspace starts with stages Lead, Qualified, Won and Lost, and starter workflows: a score of 40 qualifies the lead, a qualified lead opens a deal priced from your catalogue, and an ad lead gets a follow-up task within an hour. Change them in Pipeline › Setup › Pipeline stages, Scoring rules and Workflows.

## Where things live

| Screen | What it controls |
| --- | --- |
| Pipeline › Setup › Pipeline stages | The stages a contact moves through (starter: Lead, Qualified, Won, Lost). |
| Pipeline › Setup › Scoring rules | Points for what a contact said or did. Rules can be written in words on Pro and up. |
| Pipeline › Setup › Workflows | **When** something happens, **Then** do something: assign an owner, create a task, open a deal, move a stage, sync to your CRM, stop campaigns. |
| People › Setup › Fields you collect | The details the agent captures and the screens show. |

## The starter workflows

- **Score qualifies the lead** — at a score of 40 the lead becomes qualified.
- **Qualified lead → pipeline** — a qualified lead opens a deal.
- **Keep the deal value current** — the deal is priced from the products the customer is interested in, at catalogue prices.
- **Follow up an ad lead** — a task due in an hour for every lead from an ad.

A deal's value is always a catalogue price, never a number typed in the chat. Routing in words ("hot leads to Anuj") can be set with [Edit agent](https://www.linkubit.com/docs/edit-your-agent).

## Try a rule before it acts

- **Evaluate only.** A scoring rule or workflow has three states: **On**, **Evaluate only** and **Off**. Evaluate only checks it and counts what it would have done, but adds no points and runs no action — each row says "Would have run N times". Switch it on when the count looks right. Rules that Edit agent or an accepted suggestion creates, and any duplicate you make, start as Evaluate only. Agent rules are only On or Off: a safeguard that only watches is a safeguard that is off.
- **Backtest** in a rule's form shows, before you save, how many recent contacts the scoring rule matches, or how often the workflow would have run over the last 30 days of its trigger.
- **Test on a contact** picks one person and says whether the rule matches them now, with their value for each part of the condition — and for a scoring rule, their score before and after. Nothing is changed.
- **History**, on each row of scoring rules, workflows and agent rules, lists every change, newest first, with who made it; **What changed** shows each field before and after, and **Put back** restores a version as a new change. **Recently deleted**, beside the list's buttons, brings back a rule deleted in the last 90 days.
- When a scoring rule changes, every lead's score is worked out again in the background.

## Related
- [Qualifying leads on whatsapp](https://www.linkubit.com/guides/qualifying-leads-on-whatsapp)
- [Lead scoring](https://www.linkubit.com/features/lead-scoring)
- [Edit your AI agent in plain words](https://www.linkubit.com/docs/edit-your-agent)
