================================================================================
LAKESIDE AUTO SUPPLY - PRACTICE SAMPLE DATA
Self-Serve AI Training Toolkit
================================================================================

WHAT THIS IS

A complete, self-consistent set of sample files for a fictional independent
aftermarket parts distributor. Use it to practice AI prompts and workflows
before you point them at your own business's real numbers.

Lakeside Auto Supply, its territory managers, its accounts, its six vendor
lines and every number in these files are invented for training. None of it
describes a real company, person, or transaction. All of it is safe to paste
or upload into any AI tool, including public ones.

THE COMPANY IN ONE PARAGRAPH

Lakeside Auto Supply is an independent aftermarket parts distributor founded
in 1987, headquartered in Toledo, Ohio, with a second branch in Findlay.
About 120 active accounts across northwest Ohio -- independent repair shops,
tire dealers, a handful of small fleets, and a couple of small-town jobbers.
Six vendor lines on the line card, covering brakes, suspension, filtration,
batteries, chassis & steering, and wipers & lighting. Six outside territory
managers cover the two branches' delivery routes. Member of the Great Lakes
Parts Network buying group.

--------------------------------------------------------------------------------
THE FILES
--------------------------------------------------------------------------------

Lakeside_Distributor_Profile.txt
    The "who we are" context document -- locations, delivery routes, line
    card, customer mix, buying group, and competitive set. Paste this in
    whenever a prompt asks for company context.

Lakeside_Account_List.csv
    40 of Lakeside's ~120 active accounts: account_id, account_name, town,
    type (shop / tire dealer / fleet / jobber), territory_manager, and the
    year they first ordered from us.

Lakeside_Invoice_Export_Jan-Jun_2026.csv
    THE ground-truth file. Account x line x month revenue, January through
    June 2026, 672 rows. This is the shape of report your own invoicing or
    distribution system hands you. Every dollar figure elsewhere in this
    dataset traces back to this file.

Lakeside_Field_Notes_Jan-Jun_2026.csv
    94 territory manager field notes over the same six months -- visits,
    calls, and what shops actually said. This is the "why" layer that
    explains what the invoice file only shows as numbers.

Lakeside_Counter_Notes_Jan-Jun_2026.csv
    25 short inside-sales / counter notes -- order comments and phone-call
    fragments logged by the counter staff at both branches. Field reps don't
    always see these. That's the point of one of the exercises.

--------------------------------------------------------------------------------
SUGGESTED FIRST EXERCISES
--------------------------------------------------------------------------------

1. Account performance review. Hand the AI the invoice export alone and ask
   it to flag the accounts worth a closer look. Then add the field notes to
   the same conversation and ask it to explain the causes. Don't skip ahead
   to the field notes -- see what the numbers alone tell you first.

2. White-space check. Ask which healthy accounts (buying multiple lines
   steadily) have never bought a specific line. Cross-check the answer
   against the account list to make sure it's not just a data gap.

3. Fill-rate escalation. Search the field notes for backorder or fill-rate
   language and match it against the invoice trend on that line. Practice
   telling a supply problem apart from a selling problem.

4. Cadence versus results. Count field notes per account, then compare
   against revenue direction. Visiting a lot and selling more are not
   the same thing -- find the account where they diverge in each direction.

5. Counter-to-field gap. Read the counter notes alongside the field notes
   for the same account. Look for something the counter staff heard that
   never made it into a field visit.

6. The absence exercise. Pick an account with fading revenue and ask the AI
   what the field notes say about it. Sometimes the most important finding
   is that there's nothing there to read.

All prompts are on the workshop resources page.
Questions: lauren@tromml.com
================================================================================
