Opportunities and Pipelines

An opportunity is a potential piece of business attached to a contact. A pipeline is the ordered set of stages that opportunity moves through. Together they are how a build answers the two questions a client asks every week: what is in progress, and what is it worth.

The important design idea is that stages should describe what has happened, not what someone intends to do. Booked consultation and proposal sent are observable events that automation can detect and move the card on. Following up and interested are states of mind, and a pipeline built from those requires a human to maintain it, which means within a month it will be wrong. The best pipelines in this platform are almost entirely automated, because every stage boundary corresponds to an event a workflow can see.

Treat the pipeline as a reporting instrument that happens to look like a task board. Its value is not that reps can drag cards; it is that the client can see conversion between stages, time spent in each, and value at risk, without trusting anyone's memory. That only works if a card's position is a consequence of events rather than an act of discipline.

Pipelines and stages

• Multiple pipelines, typically one per distinct sales process or service line
• Ordered stages with clear names, plus the won and lost terminal stages
• Whether a pipeline is visible to all users or restricted to certain roles
• Stage-level expectations such as the event that qualifies an opportunity to be there

The opportunity record

• Name, linked contact, owner, value, source and status
• Expected close or appointment date
• Custom opportunity fields for service, location, product or referrer
• Lost reason, as a controlled dropdown rather than free text

Automation

• Workflow triggers on opportunity created, stage changed, value changed, status changed, or an opportunity sitting too long in a stage
• Workflow actions that create an opportunity, move it, set its value, change its owner or mark it won or lost
• Rules connecting pipeline movement to appointment status, payment events and form submissions

Ownership and routing

• Assignment rules: round robin, by territory or service, or by availability
• Reassignment on escalation or absence
• Visibility so a rep sees their own board while a manager sees everything

Reporting views

• Board view by stage, list view with filters, and value totals per stage
• Conversion between stages and average time in stage
• Filters by owner, source, date range and pipeline

1. Map the client's actual process by asking what happens, in order, from first inquiry to money received. Write it as events, not intentions.
2. Cut the list down. Five to seven stages is almost always right; a pipeline with twelve stages will not be maintained.
3. For each stage, name the single event that puts an opportunity there. If you cannot name one, the stage is a feeling and should be merged or removed.
4. Decide whether one pipeline is enough. Separate pipelines are justified by genuinely different processes, not by different services following the same steps.
5. Create the pipeline and stages with the won and lost stages at the end, and agree what won actually means with the client: contract signed, first payment, or work completed.
6. Create the lost-reason dropdown with the client's real reasons. This single field is where most of the useful analysis in the pipeline comes from, and free text destroys it.
7. Decide the value convention: total contract value, first payment, or expected monthly value. Write it down, because inconsistent values make the pipeline total meaningless.
8. Build the creation workflow so opportunities are created automatically on the qualifying event rather than by hand. Manual creation is skipped and the pipeline undercounts.
9. Build a movement workflow per stage boundary, each triggered by the event you identified. Appointment booked, appointment completed, payment received and form submitted are the usual four.
10. Set the assignment rules and test them with several contacts, including when the assigned user is unavailable.
11. Add stale-opportunity handling: a workflow that fires when a card has sat in a stage too long, creating a task for the owner rather than silently letting it age.
12. Add the won and lost automation: what happens to the contact's tags, what onboarding or nurture starts, and what reporting event is recorded.
13. Test the whole path with a test contact from inquiry to won, and again to lost, confirming the card moves without anyone dragging it.
14. Train the client on what they are allowed to change by hand, which should be very little: usually value, lost reason and nothing else.
15. Review the board with the client after a few weeks and look specifically for cards in the wrong place, which tells you which stage boundary is not automated properly.

Opportunities sit between Contacts and Reporting, and they are moved almost entirely by Workflows reacting to Calendars, Forms, Payments and Conversations. The AI agents can create and update them when a conversation reaches a qualifying point.

Lead to booked to completed

A form submission creates an opportunity at the first stage. The AI agent books an appointment, and a workflow moves the card to booked. Appointment completion moves it to the next stage and creates a follow-up task. A payment event moves it to won. Nobody touched the board, and the conversion rates between those stages are now real numbers.

No-show and recovery

An appointment marked no-show moves the opportunity back a stage and triggers a rebooking sequence. If the contact rebooks, the card moves forward again. Without this, no-shows sit in a booked stage forever and inflate the client's view of their pipeline.

Stale pipeline cleanup

A workflow watches for opportunities that have not moved within a defined period and creates a task for the owner to either progress or close them with a reason. This is what keeps a pipeline total honest, and it is the single highest-value automation you can add to an existing messy board.

• Stages built on intentions will always be stale. If a stage cannot be entered by an observable event, the pipeline becomes a manual chore and the client stops trusting it.
• Too many stages is the most common design error. Each stage boundary needs an automation and a meaning. Twelve stages means twelve things to keep correct and none of them will be.
• Manual creation undercounts everything. If reps create opportunities by hand, a meaningful share of real leads never appear, and conversion rates look better than reality.
• Inconsistent value conventions make totals meaningless. One rep entering monthly value and another entering annual value produces a pipeline number nobody can use. Decide and enforce it.
• Free-text lost reasons waste the most useful field on the record. Use a controlled list, keep it short, and review it with the client periodically.
• Duplicate opportunities. A contact who submits two forms can end up with two cards, double-counting the value. Condition the creation workflow on whether an open opportunity already exists.
• Cards do not move themselves backwards. Cancellations, no-shows and refunds need explicit automation, or the board only ever reflects the good news.
• Stage-change triggers can loop. A workflow that moves a card on stage change can re-trigger itself. Check for cycles whenever a workflow both listens to and writes stage.
• Won is ambiguous until you define it. Signed, paid and delivered are three different dates, and the client's revenue reporting depends on which one you chose.
• Multiple pipelines fragment reporting. Splitting by service when the process is identical means every report has to be assembled from several boards. Split only on genuinely different processes.
• Permissions and visibility surprise people. Reps seeing each other's deals, or a manager not seeing everything, is usually noticed at the worst moment. Test with a real non-admin user.
• Historic data does not backfill. Adding a stage or changing the pipeline does not reinterpret past opportunities, so conversion trends break at the point you changed the structure. Make structural changes deliberately and note the date.