# Dollar General put AI in 19,000 stores to fix the oldest problem in retail *The discounter automated inventory decisions across distribution and stores, cutting out-of-stocks while avoiding overbuys.* By **Jenny Huang Goodman MPA MSc MHSA, Principal** — The Stash Edge, Hako Shikin LLC. Published 2026-09-04. Canonical: https://www.pops4.com/stash/articles/dollar-general-2026-09-04t15-5 Subject: Dollar General Tags: inventory, supply chain, ai, retail, allocation, discounters --- Dollar General installed AI-driven inventory systems across its **19,000** stores and distribution network, according to Retail Dive, addressing the core physical-product problem: getting the right SKU to the right location at the right time. The system decides what ships where, when to reorder, and how much safety stock each store holds. The retailer feeds the AI real-time sales velocity, local demand patterns, supplier lead times, and shelf capacity. The system adjusts allocation store by store, replacing the legacy model where regional buyers guessed volume and every location got the same case count. Distribution centers now pack trucks based on predicted sell-through, not last year's average. This works because Dollar General operates at high SKU density in small footprints. A rural store has **200** square feet less backroom than a suburban unit, so overstock kills margin and out-of-stocks kill traffic. The AI reads which stores move seasonal faster, which demographics prefer name-brand over private label, and which zip codes spike on specific items during weather events. It reallocates before a human buyer sees the pattern. The mechanism is **dynamic safety stock**: the system raises or lowers buffer inventory by location and SKU based on variance in demand and supplier reliability. A store that historically runs out of paper towels on Fridays gets a higher threshold. A store that never sells a certain snack SKU gets none. The result is fewer emergency mid-week deliveries, less markdowns on overstock, and higher in-stock rates on fast movers. A small physical-product brand can run the same play without enterprise software. Start with a **90-day lookback** by SKU and sales channel. Export your order data into a spreadsheet. Tag each SKU with sales velocity, return rate, and supplier lead time. Create three columns: slow (moves fewer than **10 units/month**), steady (**10-50**), fast (**50+**). Assign a reorder point to each: slow SKUs get zero safety stock, steady get one week, fast get two weeks. Next, segment your customer list by order frequency and average order value. Flag accounts that reorder predictably versus those that buy sporadically. Allocate inventory to the predictable accounts first when supply is constrained. If you manufacture or import, build a simple model: lead time in days, minimum order quantity, and holding cost per unit per month. Reorder when remaining inventory equals lead time demand plus your safety stock target. Run this monthly. If you sell direct-to-consumer and wholesale, treat each channel as a separate store. Track sellthrough rate by channel. If DTC moves **5 units/day** and wholesale moves **2**, allocate new inventory **70/30** until the ratio shifts. Use a **$20/month** inventory app like Cin7 or Stocky to automate reorder alerts. The cost is negligible and the system replaces guessing. The broader pattern is that inventory intelligence is no longer a scale advantage. Dollar General spent millions, but the logic—demand variance drives buffer stock, not gut feel—applies to a **100-unit** production run. The brands that win in physical product now are those that allocate inventory like a quant fund allocates capital: by signal, not by habit. ## The takeaway AI inventory allocation beats gut feel by adjusting safety stock and reorder points based on SKU velocity and location demand variance. --- ## Publisher **Hako Shikin LLC** — Virginia Beach, Virginia. Founded 1997. ASI 217876 · DUNS 18-204-6339. Principal and author: **Jenny Huang Goodman MPA MSc MHSA**. - Author: https://www.huanggoodman.com/about - LLM context: https://www.pops4.com/stash/llms.txt - MCP endpoint, for AI agents: https://mcp.pops4.com/mcp - Client dashboard: https://dashboard.pops4.com/ - Catalogue: 70,000+ products, 200+ brands