Turn a vague 'this process is broken' complaint into a ranked bottleneck audit

You are auditing one specific internal process that people keep complaining about, to find where it actually breaks down and what's realistic to fix first.

PROCESS UNDER REVIEW
New client onboarding, from signed contract to first invoice.

CURRENT STEPS AS THEY ACTUALLY HAPPEN
Sales emails ops a PDF contract, ops manually re-keys it into the billing system, finance waits for ops to confirm before setting up the account, no one owns following up if ops is backlogged.

THE PAIN POINT PEOPLE REPORTED
Clients are waiting 9-12 days for their first invoice and sales is fielding angry calls.

SOMETHING THAT CANNOT CHANGE
The billing system cannot be replaced this year; it's mid-contract with the vendor.

WHO WOULD ACTUALLY APPROVE A FIX
The Ops Director, who controls the ops team's workflow tools and priorities.

Do this in four phases.

PHASE 1 — MAP THE REAL FLOW
Rewrite the steps as they actually happen, not as the org chart says they should happen, and number each one. Where the reported pain point doesn't map cleanly onto any single step, say so explicitly rather than forcing it onto the nearest one — a complaint about 'it's slow' often traces back to a handoff between two steps, not either step itself.

PHASE 2 — FIND THE BOTTLENECK, NOT A SYMPTOM
Identify which specific step or handoff is the actual constraint — the one where work queues up waiting on a single person, a single approval, or a single tool — and distinguish it from steps that are merely annoying but not actually slowing anything down. A step can be tedious and not be the bottleneck; only flag a step as the bottleneck if removing it would measurably speed up the whole process.

PHASE 3 — RESPECT THE FIXED CONSTRAINT
Every fix you propose must work within the stated unchangeable constraint. If the constraint itself is the bottleneck, say that plainly instead of proposing a fix that quietly assumes the constraint away — a recommendation the decision-maker can't actually implement is worse than no recommendation.

PHASE 4 — RANK REALISTIC FIXES
Propose three fixes ranked by effort-to-impact ratio, cheapest and fastest first. For each, state what it requires from the named decision-maker specifically — an approval, a tool purchase, a policy change — since a fix with no clear owner never gets implemented.

WHAT NOT TO DO
Do not recommend generic process-improvement moves ('implement better communication', 'add a project management tool') unless you tie them to the specific bottleneck found in Phase 2. Do not propose a fix that requires unlimited headcount or budget without flagging that as the trade-off.

OUTPUT FORMAT
1. Numbered flow map with the pain point located on it.
2. One paragraph naming the actual bottleneck.
3. Three ranked fixes, each with owner and rough effort.
4. One line on what happens if nothing changes.