Skip to content

E-commerce operations

Reference build

Stockouts down 61% across hundreds of SKUs

An agent that watches stock levels, sales velocity and supplier lead times across hundreds of SKUs, and drafts reorders before the shelf empties rather than after.

61%fewer stockouts

At a glance

Duration
17 weeks
Team size
1 person
Engagement
New build
Project type
AI & automation
Industry
E-commerce & Retail

The situation

The challenge

Reorder points set once go stale as velocity and lead times move. Across hundreds of SKUs nobody revisits them, so the first signal is usually a stockout.

What we did

Recompute reorder points continuously from actual velocity and observed lead times, and draft rather than execute.

The calls that mattered

  • Lead times observed, not configured

    Supplier lead time is measured from delivery history rather than taken from a field somebody filled in once.

  • The agent drafts, a human commits

    Purchase orders are money. Drafting removes the work without moving the decision.

What changed

reduction in stockouts
61%reduction in stockouts
of reorders drafted for human approval
100%of reorders drafted for human approval

Stockouts down 61%, with every reorder arriving as a draft carrying the velocity and lead-time reasoning behind it.

Services used

  • Continuous reorder-point calculation from live velocity
  • Observed supplier lead-time tracking
  • Draft purchase orders with reasoning attached

What we would do differently

Every project has one of these. Publishing it is the point — a case study with no regrets in it is marketing, not evidence.

Seasonality was the hard part. Velocity alone over-orders after a spike, so the smoothing window mattered more than anything else in the model.

Services used

More work

Ready to start?

Start a project like thisWork