Daily BigCommerce low-stock digest with restock picks

Every weekday at 8am, generate a BigCommerce low-stock digest with restock recommendations and post it to a Slack channel. This is an agent workflow (not code) because ranking, grouping, and the reorder-now-versus-watch judgement all depend on what the numbers actually say.

Trigger: cron, every weekday (Mon–Fri) at 8:00am in the store's local timezone. BigCommerce is not a poll provider, so cron is the right trigger.

Adjustable inputs to expose at the top of the workflow:

• SAFETY_THRESHOLD — the inventory level at or below which a variant counts as low stock (default: 10).

• SLACK_CHANNEL — the Slack channel name or ID where the digest should be posted.

• LOOKBACK_DAYS — how many days of sales history to use for velocity (default: 30).

Step 1 — Find the low-stock variants. Call BigCommerce Get All Variants and page through the full result set. Keep only variants where inventory_level is at or below SAFETY_THRESHOLD. Skip variants where inventory_tracking is 'none' (those have no meaningful stock count). Record sku, product_id, variant_id, inventory_level, and price for each surviving variant.

Step 2 — Enrich with product context. For each low-stock variant, use BigCommerce Get All Products (or Get a Product per unique product_id, whichever is cheaper given the count) to attach product name and primary category. Cache product lookups so the same product isn't fetched twice.

Step 3 — Pull sales velocity. Call BigCommerce Get All Orders filtered to the last LOOKBACK_DAYS days. Exclude non-sales statuses (Cancelled, Refunded, Declined, Incomplete, Pending). For each surviving order, call List Order Products to get the line items. Sum quantity by SKU across every line item to get units_sold_30d for each SKU. SKUs with no sales in the window get zero.

Step 4 — Judge each low-stock variant. Compute daily_velocity = units_sold_30d / LOOKBACK_DAYS. Compute days_of_cover = inventory_level / daily_velocity (or mark as 'no recent sales' when velocity is zero). Then bucket:

• Reorder now — the SKU is at or below threshold AND selling at a meaningful clip. Use judgement: an item with 3 days of cover left is clearly urgent; an item with 60 days of cover isn't. Look at the actual distribution of days_of_cover across the low-stock set and pick the cutoff where the ranking naturally breaks.

• Watch — below threshold but selling slowly or not at all. These don't need reordering yet; they need attention.

For each Reorder-now item, suggest a reorder quantity sized to cover roughly the next 30 days at current velocity, rounded up to a sensible whole number.

Step 5 — Draft the digest as Slack mrkdwn. Open with a one-line headline showing totals, e.g. '*12 low-stock SKUs today — 4 need reordering now*'. Then two sections:

• *Reorder now* — one bullet per item: 'Product name (SKU) — X units left, selling ~Y/day, ~Z days of cover, suggested reorder: Q units'. Sort by urgency (fewest days of cover first).

• *Watch* — one bullet per item: 'Product name (SKU) — X units left, ~Y sold in the last 30 days, category: <category>'. Sort by SKU or category for scannability.

If a section is empty, say so explicitly ('Nothing selling fast enough to worry about today'). If no variants are below threshold at all, post a single 'All inventory above threshold' message instead of the full digest.

Step 6 — Post the digest to SLACK_CHANNEL using the Slack Bot Send a Message action. Use mrkdwn formatting (single asterisks for bold, hyphens or bullets for lists).

Guardrails: if Get All Variants returns nothing, post a short flag that the catalog read came back empty rather than sending an empty digest. If the velocity fetch fails partway through, post what you have and note which parts are incomplete. Never invent SKU names, categories, or velocity numbers — only use what BigCommerce actually returns.