Ateneo futurists envision AI-powered food stalls, sari-sari stores

The Ateneo de Manila University’s Business Insights Laboratory for Development (BUILD) is looking at ways for artificial intelligence (AI) to augment and enhance—rather than replace—human labor in small businesses, which make up the bulk of the Philippine economy.

BUILD researchers Zachary Matthew Alabastro, Joseph Benjamin Ilagan, Lois Abigail To, and Jose Ramon Ilagan pay particular attention to a very ubiquitous tool of everyday business: the humble pen-and-paper logbook. Low cost, simple, and reliable, this analog solution is indispensable for keeping track of the multitude of small items that change hands throughout the day in small businesses, from shopping center food stalls to neighborhood “sari-sari” or convenience stores.

A handwritten ledger can be counted on for record-keeping even in the bustling and oftentimes hectic environment of a kitchen or backroom, where an electronic tablet might be too cumbersome or fragile to use. But despite their dependable simplicity, handwritten logs can be painstaking to tabulate and make sense of—especially when one is trying to glean insights into how to better run a business.

Meanwhile, AI lends itself perfectly to business data analysis: it makes short work of identifying which products are performing well or poorly; tracking sales trends over time; and offering recommendations on inventory, pricing, and restocking.

Understandably, however, many small business owners and workers hesitate to digitize out of concern over steep learning curves and job redundancy. But the Ateneo researchers propose a “copilot” model, in which AI complements and supports human effort.

Their study, presented at the recent Artificial Intelligence in Human-Computer Interaction Conference 2025 in Sweden, explores how optical character recognition (OCR) and large language model (LLM) technology can turn handwritten sales logs into more manageable digital data. Tested in an actual food stall at the Ateneo’s Student Enterprise Center, the researchers’ system is built with Python and uses Amazon Web Services for OCR and Anthropic’s Claude 3 Haiku LLM to interpret the handwritten logbook data.

The system allows even someone with no digital training to grasp inventory trends with ease. It scans logbook photos and uses AI to recognize products, match prices, and tabulate sales summaries. This helps businesses quickly identify bestsellers or slow-moving stock, thereby making it easier to keep up with customer demand.

The researchers’ early prototype shows moderate accuracy and offers much hope for improvement. It can also be adapted, they say, to handle other kinds of handwritten data such as inventory sheets, delivery logs, or even payroll ledgers.

Not unlike analog logbooks themselves, this AI tool is meant to be simple, affordable, and easily upgradable. As AI accuracy improves by training on more shorthand writing, local stalls can eventually gain reliable, low-cost access to business insights once reserved for larger enterprises.

SOURCE:
https://archium.ateneo.edu/build/2/