Trust, Talent, and Transformation: Inside Indonesia’s AI Revolution
Indonesia’s AI story is moving from ideas to execution. The National AI Strategy (Stranas KA 2020–2045) sets the tone: build skills, improve data and infrastructure, and use AI responsibly so it actually helps people and industry.
When models start touching customer journeys, operations and risk, trust, governance and cost stop being side issues and become core to growth.
Research on Indonesian NLP keeps finding the same thing: tools trained mainly on English or generic datasets miss Bahasa nuance, regional usage and context. Teams that train or adapt models on Indonesian data get better sentiment, intent and document results.
Against that backdrop, our panel from this year’s 51st edition of the AIBP conference and exhibition in Jakarta brings together Lily Wongso , EVP Enterprise IT Architecture, Data Management & Service Quality, PT Bank Central Asia Tbk (BCA); Kurnia Rosyada, Group Head of Enterprise Data Analytics, PT Bank Mandiri (Persero) Tbk.; Paul Soegianto , Chief of Group Digital Strategy, PT Astra International Tbk and Sigit Pratopo , Senior Vice President, Integrated Enterprise Data and Command Center, PT Pertamina (Persero) . Together, they discuss how Indonesia’s leading enterprises are scaling AI in practical, people-focused ways while managing cost, culture and governance.
BCA explains how explainable models build trust inside the bank. Bank Mandiri ties every AI project to revenue, productivity or risk. Astra shows why local models beat generic tools and how hybrid choices keep costs predictable and data safer. Pertamina’s Digital Factory turns a pipeline of use cases into measurable outcomes.
Key takeaways include:
1) Build the foundations first: hybrid where it counts
The panel is clear that infrastructure choices drive long-term cost and control. Astra warns about token-priced LLM surprises and makes the case for a pragmatic hybrid stack that mixes public cloud with on-prem or colocation for sensitive or high-volume work. BCA adds practical GPU timing and CPU/GPU allocation to keep unit costs sane during training and inference
2) Make governance your enabler, not your brake
Bank Mandiri ties every AI initiative to one of three outcomes: revenue, productivity or risk management, under a governance programme that emphasises transparency and explainability. That clarity speeds adoption instead of slowing it down.
3) Connect the data to scale the value
Pertamina shows how a “Digital Factory” approach converts a queue of use cases into measurable impact by orchestrating data, OT and IT across the group. Treating data connection as a product, not a project, lets teams industrialise wins and repeat them.
4) Localise to win, protect what matters
Astra found that imported tools often misread Bahasa Indonesia sentiment. Building local language models improved accuracy, protected sensitive information and created proprietary IP. Local context is not a nice-to-have. It is a competitive edge.
5) Talent is the engine of transformation
Reskilling and upskilling are non-negotiable. Astra is modernising processes while training a large workforce, and Pertamina’s execution model invests in internal capability so outcomes persist beyond the first deployment.
Listen to the conversation
These themes line up with our upcoming Indonesia discussions on integration and APIs, hybrid cloud and governance by design. Hear the full context, examples and hard-won lessons from the people doing the work.
Listen to the full episode on Spotify here
Interested in our upcoming activities in Indonesia, find out more here
