Act as an operations improvement lead. Build a practical improvement plan for [PROCESS NAME] used by [TEAM AND CUSTOMERS]. The improvement target is [IMPROVEMENT GOAL] by [TARGET DATE]. Work only from the information supplied below. Current state: [CURRENT WORKFLOW] Observed problems and evidence: [PROBLEMS AND DATA] Constraints: [CONSTRAINTS] Available roles and systems: [OWNERS AND RESOURCES] First, map the current process from trigger to completed output. For each step, identify the owner role, handoff, system of record, input, output, average time if supplied, and known failure mode. Distinguish symptoms from likely root causes. Use a simple five-whys analysis only where the evidence supports it; mark unverified causes as hypotheses. Produce: 1. A current-state table and a future-state table, each with numbered steps, owner roles, handoffs, controls, and escalation paths. 2. A prioritized improvement backlog with problem, likely cause, change, expected benefit, effort, risk, dependency, and owner role. 3. A 30/60/90-day implementation plan. For every action, specify one accountable owner role, due date or milestone, required decision, and completion evidence. 4. A measurement plan with baseline, target, cadence, data source, and one quality or customer-impact guardrail for each main metric. 5. A short operating procedure for the redesigned workflow, including exception handling and a weekly review cadence. Favor changes that remove rework, unclear ownership, unnecessary approvals, or manual copying before recommending new software. Do not invent volumes, costs, policies, staffing, capabilities, or root causes. Before answering, check that every future-state step has an owner and a system of record, and flag any action that depends on an unconfirmed assumption. Ask up to three clarifying questions only if a required input is missing.
Fill in
| Placeholder | What to enter | Example |
|---|---|---|
| [PROCESS NAME] | Name the workflow to improve, such as new vendor onboarding. | New vendor onboarding |
| [TEAM AND CUSTOMERS] | Identify the teams performing the work and the internal or external recipients of the output. | Procurement, finance, and security teams; internal requesters waiting to purchase services |
| [IMPROVEMENT GOAL] | State a measurable improvement target, such as reducing cycle time or error rate. | Reduce median onboarding cycle time from 18 business days to 10 while maintaining security review quality |
| [TARGET DATE] | State the date by which the improvement should be measured. | March 31 |
| [CURRENT WORKFLOW] | Describe the existing steps, systems, handoffs, approvals, and normal exceptions. | Requester emails procurement. Procurement creates a spreadsheet row, gathers a W-9 by email, forwards the vendor to security, then sends approved details to accounts payable for ERP entry. Status is tracked in the spreadsheet; urgent requests are chased in Slack. |
| [PROBLEMS AND DATA] | Provide observed delays, defects, rework, feedback, and available baseline data. | January sample of 42 requests: 16 waited more than four days for missing tax details, 11 were returned by security for incomplete data, and 9 had duplicate follow-up emails. Requesters reported they could not see status. |
| [CONSTRAINTS] | List budget, staffing, compliance, system, timing, or policy constraints. | No new procurement platform this quarter; security policy requires review for vendors with system access; finance can change the intake form but not the ERP. |
| [OWNERS AND RESOURCES] | List available roles, systems, reports, and decision makers without inventing named owners. | Procurement operations manager, procurement coordinator, security analyst, accounts payable specialist; Microsoft Forms, shared mailbox, spreadsheet, ERP, weekly vendor report. |
How to use
- Document the actual workflow from the first trigger through the completed output, including workarounds and exception paths.
- Paste baseline data from a defined period, not only complaints, so priorities can be tied to repeatable failure patterns.
- Confirm that each recommended owner has authority to complete the action or escalate it to the right decision maker.
- Then send: “Convert the future-state workflow into a one-page SOP for a new coordinator, including the exact information to collect at intake.”
Variations
Root cause workshop
Use when the team needs to investigate before choosing fixes.
Facilitate a 60-minute root-cause workshop for [PROCESS PROBLEM] using [EVIDENCE]. Participants are [PARTICIPANTS]. Produce an agenda, a current-state fact list, 10 neutral questions, and a five-whys worksheet that separates verified facts from hypotheses. End with a ranked list of data to collect and experiments to run before selecting solutions. Do not label people as the cause or assume a process failure without evidence. Include a decision owner role for the next step.
SOP rewrite
Use after a new process has been decided.
Write a standard operating procedure for [PROCESS] using this approved future-state workflow: [WORKFLOW]. The audience is [ROLE]. Include purpose, scope, prerequisites, numbered steps, system-of-record rules, decision points, exception paths, escalation contacts by role, and completion criteria. Add a short quality checklist that can be used before handoff. Do not change approved policy or create missing controls; flag gaps requiring an owner decision. Keep the main procedure under 900 words.
Executive improvement memo
Use when requesting leadership approval for a process change.
Write a 500-word process improvement decision memo for [AUDIENCE] about [PROCESS]. Use [CURRENT PERFORMANCE], [PROPOSED CHANGES], [COSTS OR CAPACITY], and [RISKS]. Include the problem, evidence, recommendation, expected benefit, implementation milestones, decision required, and metrics. Distinguish estimates from measured facts. Do not claim savings, compliance improvements, or customer impact unless the supplied information supports them. Flag dependencies that could delay the plan.
Tips
- Measure elapsed time and touch time separately; a workflow may have little work in it but still create long customer waits at approvals or queues.
- Map the system of record at every handoff, because duplicate spreadsheets and inbox tracking often create both rework and conflicting status.
- Use a named accountable role for each action, even when several teams contribute; shared ownership is frequently a sign that escalation will stall.
- Pilot a changed intake form or routing rule with a defined cohort before changing the process for everyone, then compare the pilot to the same baseline measure.
FAQ
Can AI identify the root cause of a process problem?
It can organize evidence and generate hypotheses, but it cannot establish a cause without operational data and observation. Test the hypothesis before investing in a solution.
What should a process improvement plan include?
It needs the current workflow, target outcome, prioritized changes, accountable owners, dates, dependencies, and a measure that proves whether the change helped.
Should process improvement start with automation?
Usually no. Remove unclear handoffs, duplicate entry, and unnecessary approvals first; automation can preserve a flawed process if the underlying work is not understood.