PMD Solutions
Quick Start12 min

Process audit checklist before you automate

Map handoffs, bottlenecks, and duplicate data entry so automation fixes the right work — not bad process. Includes a half-day agenda and deliverables checklist.

Automation doesn’t fix a broken process — it scales it. Before Zapier, Make, HubSpot workflows, or custom code, spend one focused session understanding how work actually moves. This checklist is that session.

Pick one workflow, not the whole company

Choose a path that hurts weekly: lead → first meeting, quote → invoice, onboarding → first value, support → resolution. One path. End to end.

  • Name the trigger (what starts it) and the done state (what “finished” means)
  • List every human role that touches it — including the people who “just check Slack”
  • List every tool involved (CRM, email, sheets, ERP, chat, e-sign)
  • Note where data is retyped from one system into another

Map the handoffs

Handoffs are where deals stall and blame starts. For each step, write: who owns it, what they need to start, what they produce, and who waits next.

  • Unowned steps — “someone usually handles it” means no one owns the SLA
  • Invisible waits — waiting on approval with no due date or ping
  • Tribal knowledge — only one person knows the exception path
  • Shadow trackers — side spreadsheets that contradict the CRM

Find the bottlenecks worth automating

Not every annoyance deserves a workflow. Score each step: frequency, time cost, error rate, and emotional load. Automate high-frequency, rules-based work first.

  • High volume + clear rules → strong automation candidate
  • Rare + judgment-heavy → keep human; maybe add a checklist or template
  • Errors that create rework downstream → fix data quality before speed
  • Approvals that block revenue → shorten the path or add SLAs + reminders

Decide what “good” looks like

Before building, write the success metric in one line: “Leads assigned within 15 minutes,” “Quotes sent same day,” “Onboarding checklist complete before kickoff.” If you can’t measure it, you can’t tell if automation helped.

  • Baseline the current time or error rate (even a rough sample of 10 cases)
  • Agree who gets notified when the automation fails
  • Define the rollback: how do humans complete the step if the tool is down?

Only then choose tools

Native CRM automation, iPaaS (Make/Zapier/n8n), or custom APIs — pick after the process is clear. Tool choice is secondary to knowing which handoff you’re removing.

Run a half-day audit agenda

If you need a repeatable format, use this 3–4 hour block with the people who actually touch the work (not only leadership).

  • 30 min — Agree on the one workflow, trigger, and done state
  • 60 min — Walk a real case end-to-end; note tools, waits, and retypes
  • 45 min — Score bottlenecks (frequency × time × error × emotional load)
  • 30 min — Rank: fix process, automate, or leave alone
  • 30 min — Write success metrics, owners, and failure alerts

Outputs you should leave with

  • One-page process map (roles, tools, handoffs)
  • System-of-record decisions for the facts that currently live in three places
  • Top 3 automation candidates with owners and success metrics
  • Explicit “don’t automate yet” pile — judgment-heavy or rare steps
  • A named person who owns monitoring when automation fails

When the audit is done, you should have: a one-page map, a short list of automations ranked by impact, and a clear “don’t automate yet” pile. That’s enough to build without regret.

Need a second pair of eyes on the process?

Bring us one workflow. We’ll map the handoffs, flag what to automate, and what to fix first.