Packworks reports AI lifts sari-sari store sales

According to a report by Filipino tech startup Packworks, sari-sari stores using the company's AI-powered Store Insighting Project (SIP) posted large short-term gains. Packworks reports that an analysis of more than 1 million monthly transactions across its network of 300,000 micro-retailers found a 79% increase in median sales for the top 50 products identified by SIP. The report states that median gross merchandise value (GMV) per store rose from P187,229 to P335,818, seasonal-product GMV rose 47%, and slower-moving-product GMV rose 96%, from P7,361.60 to P14,429. Packworks also reports an overall 29% increase in median total sales and a 20% increase in median transactions for stores using SIP.

Packworks describes SIP as a personalized report that converts each store's transaction history into stocking and promotional recommendations. Per Packworks, SIP was developed with support from the Department of Science and Technology-Philippine Council for Industry, Energy and Emerging Technology Research and Development (PCIEERD). The sources do not disclose model architecture, training data beyond the transaction counts, or any evaluation methodology beyond the aggregate statistics cited in the report.

Industry-pattern observations: AI-driven inventory recommendations commonly use transaction-level sales history, seasonality signals, and local demand features to rank high-turn products and identify slow movers. Projects that personalize recommendations at the store level often produce outsized percentage gains for small merchants because small baselines amplify relative improvements. Observers should also expect selection effects: measured participants may be early adopters or more digitally engaged, which can bias results upward compared with the full population of stores.

Editorial analysis: For micro-retail ecosystems such as Philippine sari-sari stores, tools that convert transaction records into operational guidance can change stocking behavior quickly. Packworks frames the SIP findings as evidence that democratized data can boost grassroots entrepreneurship; that framing is present in direct quotes from Packworks executives cited in the report. Multiple outlets reproduced Packworks' reported metrics and store-owner anecdotes, indicating consistent messaging but not independent verification of causal attribution beyond the internal analytics described.

Industry observers and practitioners should look for:

• •published methodology or a technical appendix from Packworks that details sample selection, statistical controls, and time windows
• •replication or third-party audits that validate causal impact versus correlated adoption effects
• •product-level rollout metrics, such as retention and repeat uptake among less digitally active stores. Packworks has included direct quotes from participating store owners in the report, and Packworks' collaboration with PCIEERD is reported in the sources, but the public materials cited do not include raw data or model disclosures