AI at BUILD

Reliable AI workflows, LLM tutors and simulators, and AI for small businesses

BUILD News

13 Oct 2026

PSA to run Ateneo-built National ID verification for local governments and NGOs

Built for areas with limited or no connectivity, including geographically isolated and disadvantaged areas. The same software is available to banks and other institutions.

03 Jun 2026

Philippines among top 3 nations in AI Olympiad hosted by Ateneo BUILD

The Philippine delegation was selected through the International Olympiad in Artificial Intelligence Philippines (IOAIPH), hosted by the Ateneo de Manila University Business Insights Laboratory for Development (BUILD) of the John Gokongwei School of Management.

Ateneo BUILD › AI at BUILD

AI at BUILD

BUILD (Business Insights Laboratory for Development) is a research and innovation lab of the Ateneo John Gokongwei School of Management. Its teams build tools with large language models (LLMs), test them with real users, and publish the results.

Reliable AI workflows  ·  LLM tutors and simulators  ·  AI for small businesses  ·  GenAI for social good · FAQ

Reliable AI workflows

How to make LLM tools dependable in everyday operations. BUILD's approach, architecture-driven determinism, replaces open-ended prompting with fixed, scripted steps and built-in checks, so that AI-assisted work is predictable and a person approves what goes out.

Publications and preprints (7)
  • Ilagan, J. B., Uy, G. A., & Ilagan, J. R. (2026). Draft only, never send: Hardening AI-assisted workflows in a university research lab. IEEE Software (early access). doi.org/10.1109/MS.2026.3716868
  • Ilagan, J. B., & Ilagan, J. R. (2026). A constraint-driven conversational architecture for staged task interaction using large language models. International Journal of Advanced Computer Science and Applications, 17(2). doi.org/10.14569/IJACSA.2026.0170207
  • Alabastro, Z. M., To, L. A., & Ilagan, J. B. (2026). Where autonomy ends: Evaluating agentic LLM reliability in educational administrative workflows. ICCE 2026. Accepted.
  • Ibarra, N. M., & Ilagan, J. B. (2026). Taglish-aware AAC with constrained large language models: System design and pilot expert evaluation. HCI International 2026 Late Breaking Work (LNCS 17357). Springer. In press.
  • Ilagan, J. R., & Ilagan, J. B. (2026). Cost and capacity measurements for self-hosted agentic coding workshops in under-resourced settings. ICCE 2026 Workshop on ICT Trends in Emerging Economies. Accepted.
  • Ilagan, J. B., Uy, G. A., & Ilagan, J. R. (2026). From prompt to pipeline: Why and how AI-assisted workflows harden into deterministic systems. Preprint, SSRN. ssrn.com/abstract=6358278
  • Ilagan, J. B., & Ilagan, J. R. (2026). Operating an LLM-assisted multi-mode extraction pipeline in practice: An experience report across four heterogeneous public-data cases. Preprint, SSRN. ssrn.com/abstract=6366020

LLM tutors and simulators for management and entrepreneurship education

Conversational tutors and simulators that let students practise customer interviews, customer validation and startup financial modeling. Teachers can build new simulators by describing them in plain language. LiteChat, BUILD's classroom simulation platform, runs these simulators on several LLMs.

Featured by OpenAI

BUILD's recent talks and panels are listed on the BUILD page.

Publications and preprints (8)
  • Uy, G. A., Ilagan, J. B., Ilagan, J. R., & De Asis, F. (2026). AI-driven adaptive simulations for management education: Student perceptions from classroom deployments in Philippine higher education. HCI International 2026 Posters (CCIS 3051, pp. 153–162). Springer. doi.org/10.1007/978-3-032-30833-7_17
  • Uy, G. A., Ilagan, J. B., Ilagan, J. R., & De Asis, F. (2026). Teachers as vibe-coders: Authoring adaptive learning simulators through natural language. Adaptive Instructional Systems, HCII 2026 (LNCS 16737, pp. 558–568). Springer. doi.org/10.1007/978-3-032-30015-7_35
  • Ilagan, J. B., Uy, G. A., Ilagan, J. R., & De Asis, F. (2026). From instructor to author: Vibe coding as AI-mediated instructional design in higher education. ICCE 2026. Accepted.
  • Alabastro, Z. M., & Ilagan, J. B. (2026). Large language models and jobs-to-be-done framework in customer validation for product managers. ICICT 2026 (LNNS 2116, pp. 492–502). Springer. doi.org/10.1007/978-3-032-36012-0_40
  • Ilagan, J. B., Ilagan, J. R., To, L. A., & Alabastro, Z. M. (2025). An adaptive simulated startup financial modeling mentor using a large language model to address shortages in skilled advisors. Adaptive Instructional Systems, HCII 2025 (LNCS 15812, pp. 56–66). Springer. doi.org/10.1007/978-3-031-92967-0_4
  • To, L. A., Alabastro, Z. M., & Ilagan, J. B. (2025). Evaluation of a customer development simulator using large language model-based multi-agent system and synthetic respondents. ICICT 2025 (LNNS 1441, pp. 261–270). Springer. doi.org/10.1007/978-981-96-6938-7_22
  • Ilagan, J. B., Ilagan, J. R., To, L. A., Uy, G. A., & Rodrigo, M. M. T. (2026). Design-based refinement of an LLM-powered conversational learning simulation for customer validation in entrepreneurship education. Preprint, SSRN. ssrn.com/abstract=6520018
  • Ilagan, J. B. (2025). Reclaiming human learning in the age of generative AI: Reflections from technology entrepreneurship and higher education in the Philippines. Preprint, SSRN. ssrn.com/abstract=5363301

How do AI models play customers?
Early-stage study presented at ICICT 2026.

Can AI playing experts help design a learning app?
Early-stage study presented at ICCE 2026.

AI for small businesses with little digital adoption

Low-cost AI tools for sari-sari stores, food stalls and other micro and small businesses that still keep records on paper. KLARO, a student venture that started at BUILD, reads handwritten store logbooks with a phone camera and an LLM.

BUILD has also worked with Packworks, a Philippine technology company that serves sari-sari stores, on a data warehouse and analytics for its AI tools for store owners. Coverage: Philstar, December 2025 and Manila Bulletin, June 2026. In that work, store data stays in one environment and the transformations run in another, with personal data removed. Joben Ilagan discussed this setup at the Packworks-hosted roundtable on sovereign AI during Philippine Startup Week 2026.

Publications (2)
  • Alabastro, Z. M., Lui, B. A., Ilagan, J. B., Ilagan, J. R., & Uy, G. A. (2026). AI-powered enterprise system for CRM and inventory management in low-digital MSMEs. HCI in Business, Government and Organizations, HCII 2026 (LNCS 16729, pp. 3–21). Springer. doi.org/10.1007/978-3-032-30044-7_1
  • Alabastro, Z. M., Ilagan, J. B., To, L. A., & Ilagan, J. R. (2025). Applied optical character recognition and large language models in augmenting manual business processes for data analytics in traditional small businesses with minimal digital adoption. Artificial Intelligence in HCI, HCII 2025 (LNAI 15822, pp. 267–276). Springer. doi.org/10.1007/978-3-031-93429-2_18

Can a low-cost app read a food stall’s notebook?
Early-stage study presented at HCII 2025.

GenAI for social good: bullying prevention

BUILD works with other Ateneo units on using generative AI to create materials that help prevent bullying. Partners include the Department of Psychology’s Public Mental Health Lab, the Department of Communication, the Ateneo Laboratory for the Learning Sciences (ALLS), the Eugenio Lopez Jr. Center for Multimedia Communication (ELJ-CMC), Areté, and the Developing Evidence-Informed Digital Materials to Prevent Bullying Sandbox Residency.

Timeline and related news
  • 15 February 2025: GenAI for Social Good Workshop at Areté. Teams of college students used GenAI to make materials that raise awareness about bullying, including a website that helps parents spot and address bullying, card games for children, and a branching-story book that shows bullying from several points of view.
  • 16 May 2025: GenAI Camp on Bullying Prevention, led by the Public Mental Health Lab, with input from DepEd’s Bureau of Learner Support Services and Bureau of Learning Protection. Nineteen students from Psychology, Computer Science, Management, Information Design and Communication Technology at Ateneo and CIIT heard from Grade 7 public school students and their parents, then pitched digital bullying-prevention materials. Joben Ilagan of BUILD taught GenAI and prompt architecture.

Related news

Research team

  • Joseph Benjamin R. Ilagan, PhD, Director of BUILD and Assistant Professor, Department of Quantitative Methods and Information Technology
  • Gabrielle Anne L. Uy
  • Jose Ramon S. Ilagan
  • Zachary Matthew Alabastro
  • Lois Abigail To
  • Nirel Marie Ibarra

Frequently asked questions

Which Ateneo lab researches reliable AI workflows?

The Ateneo Business Insights Laboratory for Development (BUILD) of the John Gokongwei School of Management. Its research on reliable AI workflows studies how to make large language model tools dependable in everyday operations, and is published in IEEE Software and other venues.

Can a company or government agency work with BUILD on an AI project?

Yes. Companies, government agencies, local governments and NGOs can bring an AI workflow problem to a BUILD faculty–student team, or ask about workshops. Email build.jgsom@ateneo.edu.

Can Ateneo students join BUILD's AI projects?

Ateneo students from any program who want to work on BUILD's AI projects can email build.jgsom@ateneo.edu.

What was BUILD's role in the International Olympiad in Artificial Intelligence?

BUILD founded the Philippines' participation in the International Olympiad in Artificial Intelligence and was its accredited national coordinator for the 2026 season. BUILD ran the national selection: 226 students went through five stages, and BUILD hosted the national finals. The finalists formed Team Philippines. At the Asia-Pacific Olympiad in AI on 13 June 2026, the team tied 6th of 18 countries, with 1 gold, 1 silver, 2 bronze and 2 honourable mentions. At IOAI 2026 in Astana, Kazakhstan, all four members of the IOAI team won medals, 1 silver and 3 bronze. They competed against 466 students from 108 countries and territories. From the 2027 season, AI Olympiad Society Philippines (AIOS PH) is the accredited coordinator. IOAI Philippines (IOAI PH) is a joint project of AIOS PH and Ateneo BUILD, and BUILD has been its official host since 2026. The full IOAI history is on the BUILD page, under Programs.

Contact

Email build.jgsom@ateneo.edu. More on BUILD's other programs, including National ID verification and student venture incubation, is on the BUILD page. More research videos are on the Ateneo BUILD YouTube channel.

Back to Ateneo BUILD

BUILD News

13 Oct 2026

PSA to run Ateneo-built National ID verification for local governments and NGOs

Built for areas with limited or no connectivity, including geographically isolated and disadvantaged areas. The same software is available to banks and other institutions.

03 Jun 2026

Philippines among top 3 nations in AI Olympiad hosted by Ateneo BUILD

The Philippine delegation was selected through the International Olympiad in Artificial Intelligence Philippines (IOAIPH), hosted by the Ateneo de Manila University Business Insights Laboratory for Development (BUILD) of the John Gokongwei School of Management.


Events