Applied AI for Micro-Retailers: An Industry-Academe-Government Initiative
09 Dec 2025
Across Philippine neighborhoods, sari-sari stores keep daily life moving. Many run on thin margins and gut feel. Packworks, with support from government and academe, is changing that by putting clear, data-backed guidance in the hands of micro-retailers.
A recent analysis of 300+ stores found that owners who reviewed Packworks' Store Insighting Project (SIP) reports saw higher daily gross merchandise value and overall sales, even as some operated fewer selling days. Simple, accessible insights are helping store owners make faster, smarter calls.
Why this matters to government
- Micro-retail holds a large share of everyday spending. Better decisions at store level can lift local incomes without heavy subsidies.
- Data from real transactions improves visibility into demand, prices, and supply bottlenecks-useful for MSME policy, inflation response, and targeting support.
- Shared infrastructure and standards reduce duplication across agencies and programs.
What Packworks built-with DOST and Ateneo BUILD
Packworks grew a nationwide dataset from its sales and inventory platform. With the Department of Science and Technology's Philippine Council for Industry, Energy and Emerging Technology Research and Development (DOST-PCIEERD) Startup Grant Fund, the team developed precision marketing and insight-generation models that now inform SIP guidance.
Ateneo's Business Insights Laboratory for Development (BUILD) built the data warehouse and business intelligence systems that organized Packworks' data. That groundwork made the AI models possible-and scalable.
Packworks Chief Data Officer Andres Montiel put it plainly: "Even at this early stage of adoption, we've recorded increased sales and enhanced operational efficiency from stores by using the AI tools we've developed with support from DOST and through our collaborations with STT GDC and Ateneo BUILD."
He added: "As stores learn to leverage the recommendations from the AI-driven insights they can access through SIP, micro-retailers can make smarter decisions that translate into higher sales and more efficient operations."
Shared infrastructure that speeds up delivery
Through the ST Telemedia Global Data Centers (STT GDC) AI Synergy Lab, Packworks accessed the compute needed to train and run more advanced models. BUILD's earlier data engineering work ensured the data feeding those models was clean, structured, and ready.
BUILD's expanding work on public-interest data systems
Beyond retail, BUILD partnered with Co-Develop to launch the Technical Resource Lab (TRL), an initiative promoting standards-based digital public infrastructure (DPI) in the Philippines-covering digital identity, verifiable credentials, and interoperable data exchange. The goal: help organizations adopt building blocks that make programs more efficient and accountable.
BUILD also convenes Breathe Metro Manila with private firms, civic groups, and researchers to broaden access to air-quality data. Partners include Clarity Movement and scientific and health institutions such as the Manila Observatory and the Ateneo School of Medicine and Public Health Center for Research and Innovation (ACRI). These collaborations show a repeatable pattern: build shared data assets, then create decision tools that serve the public.
What agencies can do next
- Expand grant programs for applied AI that target MSME outcomes (sales growth, stockouts reduced), with clear privacy and consent guardrails.
- Back shared data pipelines for anonymized, standards-based retail indicators that LGUs and national agencies can use for policy and targeting.
- Fund digital extension services to onboard store owners-training, help desks, and micro-incentives tied to tool usage and reporting.
- Broker access to compute through public-private labs (similar to STT GDC) so startups can train models without heavy capex.
- Use DPI components-digital identity and verifiable credentials-for benefits, vendor verification, and procurement, reducing fraud and paperwork.
- Adopt outcome-based procurement for analytics tools. Start with pilots, measure rigorously, scale what works.
- Standardize impact measurement: daily GMV, sell-through rates, days of operation, and inventory turnover. Encourage randomized or phased rollouts to prove causality.
Results you can build on
Packworks continues to refine store-level insights so owners can track inventory, spot trends, and adjust quickly. BUILD's early systems work keeps those insights reliable at scale. With DOST support, this collaboration offers a simple lesson: when government, academia, and industry share the load, useful tools reach small businesses faster-and at scale.
If your team is building AI capacity for public programs, see practical learning paths by job role here: AI courses by job.
Your membership also unlocks:
