Operational Drag in Business: A 30-Minute Diagnostic

Operational drag is the avoidable time and effort between starting a business task and finishing an accepted result. It can appear as a proposal waiting for a price, a customer request handed between teams, or a report rebuilt after someone discovers a bad source. A tool may shorten one step while the whole process remains slow. The useful unit of analysis is the completed workflow, including review and correction.

Choose one workflow and one definition of done

Pick a recurring task with a clear start and a visible finished artifact: a sent proposal, a resolved ticket, an approved invoice, or a scheduled service visit. Write down the start trigger, final owner, and acceptance standard. Avoid auditing “operations” as a whole; the causes become too vague to act on.

Trace five recent cases

For each of five completed examples, record the start date, every handoff, every wait, every correction, and the finish date. Ask the people who did the work where they had to chase information or a decision. Record the time spent doing work separately from time spent waiting. A long cycle with little active work points to a different remedy than a short cycle filled with rework.

Sort the friction before fixing it

Signal	Question to ask	Possible response
Missing input	Which fact was absent when work began?	Change the intake form or require a source document.
Approval wait	Who can decide, and when are they available?	Define authority, threshold, and a backup approver.
Handoff loss	What information had to be requested again?	Use a shared handoff record with an owner.
Rework	What failed review, and why?	Put a check at the earliest point that catches the error.
Tool friction	Where is the same fact entered twice?	Remove duplicate entry or connect approved systems.

These are hypotheses until the five cases support them. If two examples stalled because pricing authority was unclear, fixing that rule may matter more than buying a new writing tool.

Use AI only where it changes the full result

AI can draft from approved inputs, classify routine requests, or summarize a source set. It cannot authorize a price, accept a legal obligation, or make an undocumented handoff disappear. Test one narrow use against a baseline: preparation time, review time, correction rate, and total cycle time per accepted result. If drafting gets faster but approvals remain the bottleneck, solve the approval path first.

Illustrative test: a proposal takes four days from discovery notes to client delivery. The team spends 25 minutes drafting but waits three days for missing scope and pricing approval. A ChatGPT draft may save ten minutes; a required scope checklist and a named pricing approver may change the delivery time more. Measure the complete cycle before and after.

Turn the audit into an owner-assigned experiment

Choose one change, one owner, and a review date. Record the baseline median cycle time, the number of corrections, and the percentage of cases finished by the promised date. Run the new process on the next five to ten cases. Keep the change only if accepted work improves without shifting hidden effort onto customers or another team.

For a company-wide prioritization approach, use the AI business strategy guide. For a specific drafting task, the client proposal workflow includes a prompt and review gate. The systems playbook addresses broader delegation after the bottleneck is identified.