Local File Workflow Intake (Onboarding)

Local File Workflow Intake (Onboarding)

1. 01

   Inventory the client's local file and data workflows before proposing any automation

   Walk through the actual folders on the machine that does the work. Log each recurring task with its frequency and the person who owns it, so the scope is grounded in observed behavior rather than a stakeholder's recollection.

2. 02

   Time each candidate task across a full week to establish a billable-hours baseline

   A rules-based sorter that clears a downloads folder in seconds looks impressive until you learn the task consumed 20 minutes a month. Capture minutes per occurrence and occurrences per month before quoting any retainer adjustment.

3. 03

   Confirm the operating system and machine ownership for every endpoint in scope

   Desktop utilities are platform-bound, and a Windows-only rule engine will not run on a Mac studio. Note whether machines are agency-managed or client-issued, since client-issued hardware often blocks installs.

4. 04

   Classify each workflow as internal margin work or client-facing deliverable

   Back-office file organization usually converts to margin; anything a client sees or audits should be priced as a value-add. Mixing the two in one proposal muddies the retainer conversation later.

5. 05

   Run a dry-run or preview pass on a copied folder before touching live data

   Preview modes that cap at 1,000 files per run are common in this category, so split large directories into batches and verify the rule logic against a sample before the first live execution.

6. 06

   Document a manual fallback for every automated step

   Desktop tools in this category are single-purpose and thin on support. Write down the two-minute manual procedure so delivery does not stall if the app breaks after an OS update.

7. 07

   Record the baseline, scope, and fallback in the client file before build starts

   This record becomes the evidence base for the first ROI review and protects the agency if the client later questions what was actually automated.