Top 10 AI Prompts and Use Cases and in the Government Industry in Indonesia

Indonesia's National AI Strategy (2020–2045) puts artificial intelligence at the heart of the country's 2045 vision - targeting healthcare, bureaucratic reform, education, food security and smart cities - while stressing ethics, talent, infrastructure and industrial research (Indonesia's National AI Strategy).

The government's national AI roadmap sharpens this into near‑term action: produce 100,000 AI talents annually and make 20 million citizens AI‑literate by 2029, with immediate priorities on public services and healthcare (Indonesia national AI roadmap and priorities).

Turning strategy into service delivery means practical reskilling - workplace programs that teach prompt writing, tool use, and applied AI can help civil servants and local teams convert pilots into scaled public benefit; one example is Nucamp's AI Essentials for Work, a 15‑week course on applied AI for any workplace (AI Essentials for Work bootcamp), a straightforward step toward building the talent pipeline needed to operationalize those national ambitions.

Bootcamp	Length	Cost (early bird / after)	Register
AI Essentials for Work	15 Weeks	$3,582 / $3,942	Register for AI Essentials for Work

“whichever country ‘controls AI can potentially control the world'.”

The methodology for this review triangulated three kinds of evidence: forward‑looking risk analysis, official policy consultations, and comparative trackers to create practical, government‑ready criteria.

A recent preprint was used to surface four systemic risk categories - disinformation, social bias, opacity in decision‑making, and privacy violations - and to prioritize safeguards such as algorithm audits and public participation.

AI Governance in the Indonesian Public Sector preprint

Ministerial signals and stakeholder input were sourced from Kemenkominfo's public consultation process around the national roadmap and ethics guidelines, ensuring the criteria reflect current regulatory direction and civil‑society concerns (Indonesia public consultations on the national AI roadmap and ethics guidelines).

Finally, a global policy tracker helped benchmark Indonesia's authorities - BRIN (Agency for the Assessment and Application of Technology), Kemenkominfo, and ministries responsible for bureaucratic reform such as KemenpanRB - against international best practices so each use case is assessed for risk level, data governance, transparency, and scalability (IAPP Global AI Law and Policy Tracker).

The result: a short list of prompts and pilots that favor demonstrable public benefit, auditability, and low‑regret deployment - because a single biased model can shape thousands of citizen outcomes in a day.

Sahabat‑AI is emerging as a practical citizen‑service tool for Indonesia by pairing a locally tuned large‑language model with easy access points that can cut friction in KTP applications, document renewals, taxation queries and other routine bureaucratic tasks; the open‑source project and its multilingual chat service (available via sahabat-ai.com) are deliberately built to understand Bahasa and regional languages so answers feel native rather than translated (Twimbit article on Sahabat‑AI's role in Indonesia).

Backed by Indosat, GoTo and NVIDIA and integrated into consumer touchpoints like the GoPay “Popular Services” tab, Sahabat‑AI's 70‑billion‑parameter model and local hosting on GPU Merdeka make secure, scalable B2G chatbots and voice assistants realistic tools for reducing trips to government counters and speeding everyday citizen transactions (Light Reading report on Indonesia's upgraded Sahabat‑AI multilingual chat service).

Feature	Detail
Model	Open‑source LLM, 70 billion parameters
Languages	Bahasa Indonesia, Javanese, Sundanese, Balinese, Batak
Access	sahabat-ai.com; GoPay “Popular Services”
Partners / Infra	Indosat, GoTo, NVIDIA, AI Singapore, Tech Mahindra; GPU Merdeka

“The new chat service, which uses Sahabat-AI's 70-billion-parameter model, is a major leap forward in developing a uniquely Indonesian AI ecosystem. Its multilingual capability, combined with enhanced accuracy, enables Sahabat-AI to better serve the diverse needs of people and businesses across the country. We've created a platform that is smarter, faster, and more affordable. Making the chat service available on the GoPay app has widened its reach to millions of people across Indonesia.”

NEXMEDIS offers a cloud‑based, modular health information system built to bring clinical decision support and EMR workflows into Indonesia's clinics and pharmacies, turning scattered patient records into a single, AI‑assisted view that helps clinicians make faster, more auditable decisions; the platform's AI‑powered EMR, pharmacy inventory controls with automatic potential drug‑interaction notifications, and real‑time queue management mean a busy public clinic can reconfigure patient flow on the fly rather than rely on paper trails.

Designed as an affordable SaaS stack, it promises scalability for regional health networks and already reports deployments across multiple cities and facilities - useful for district health offices seeking low‑friction digital upgrades.

See NEXMEDIS' product overview for a demo and feature list, or read about broader AI‑based clinical decision support services and why CDS tools can improve modern care delivery in practice (NEXMEDIS integrated health information system, AI‑based clinical decision support services, how clinical decision support tools support modern care delivery).

Feature	Detail
AI‑powered EMR	Secure patient tracking, analysis, and CDS integration
Pharmacy Information System	Inventory management + AI alerts for potential drug interactions
Queue & Patient Flow	Real‑time monitoring to adjust queues and patient throughput
Modular SaaS	Scalable clinic, pharmacy, lab, and hospital modules
Reach	100+ clinicians, 1000+ patients, 6 cities, 7 medical facilities

eFishery Mas Ahya can act as a practical advisory bridge between Jakarta's data streams and the small‑scale fisher who still leaves port at dawn: by packaging BMKG's smartphone‑accessible forecasts and INA‑WIS sea conditions into push alerts and combining those with early‑warning calendars for pond and lake pollution, advisory prompts could tell fishers when to delay a trip, move cages, or harvest early to avoid low‑oxygen die‑offs - measures already shown to cut risk in Lake Toba and other hotspots (BMKG smartphone weather and fish-distribution services (Kompas), Predictive lake pollution early-warning calendar and system (Mongabay 2018)).

Layering in proven vessel monitoring and low‑cost satellite comms - what pilots called “VMS+” and a planned 20‑nanosatellite constellation - would let advisory prompts reflect vessel location and wider fleet behavior, improving safety and helping managers spot shifts in stock distribution as seas warm during El Niño cycles; the result is not glamour but a tangible, life‑saving nudge: avoid the reef today, move cages tomorrow, harvest before the alert.

Capability	Research example
Weather & fish‑distribution alerts	BMKG smartphone services / INA‑WIS (Kompas)
Early‑warning for pollution	Predictive lake calendar & management guidance (Mongabay 2018)
Vessel tracking & satellite monitoring	VMS+ pilots (Inmarsat) and 20‑nanosatellite plan (Mongabay 2024)

“Thanks to the existence of such tool, fishermen in our group have had zero incidents of drifting or death due to bad weather at sea. Technology and fishermen's compliance awareness are the main keys.”

Nodeflux's smart‑traffic agenda brings proven adaptive signal control ideas into an Indonesian context: by fusing camera and sensor feeds with AI to tune lights every few minutes, intersections stop running on rigid schedules and start responding to real conditions - reducing stops, lowering emissions, and smoothing corridors into a true “green wave.” Real‑world pilots abroad show big dividends (corridor travel times have fallen by 36–43% in some U.S. examples), and targeted features such as AI‑based bus signal priority can cut transit trip times by more than half on key routes, a powerful lever for cities trying to make buses competitive again (Adaptive signal timing and sensor fusion technology, Adaptive Signal Control technologies case studies and examples).

Implementation needs durable detectors, reliable communications and operational capacity - issues the FHWA flags as essential for scaling - but done right the payoff is tangible: fewer idling cars, faster emergency response, and intersections that actually protect pedestrians and cyclists as well as drivers (FHWA guidance on Adaptive Signal Control Technologies (ASCT)), a practical, near‑term win for Jakarta and other crowded Indonesian corridors.

West Java's next frontier in public‑health AI blends BRIN's lab‑grade advances - an AI‑based diagnostic system that helps health workers identify malaria parasites - with recent machine‑learning work on early outbreak detection, creating a practical surveillance loop that turns clinical signals into forecasts and timely action; BRIN's diagnostic tools can sharpen case detection at the clinic level (BRIN develops AI potential in malaria diagnosis), while a feature‑based time‑series classification framework offers a replicable method to forecast outbreaks and non‑outbreaks from incidence data (Feature-based time-series classification for early outbreak detection (PLOS Computational Biology)).

In practice, stitching these approaches together could mean puskesmas and district health offices receive earlier, evidence‑ranked alerts so resources - testing, vector control, community outreach - reach high‑risk areas before caseloads surge, a concrete, low‑regret win for provincial health security.

Organization	Contact / Detail
BRIN (National Research and Innovation Agency)	Gedung B.J. Habibie, Jl. M.H. Thamrin No. 8, Jakarta Pusat 10340
Whatsapp	+62811-1933-3639
Email	ppid@brin.go.id
Copyright	© 2025 BRIN

Bank Rakyat Indonesia (BRI) can dramatically reduce payment losses and protect fragile customer relationships by adopting layered fraud risk‑scoring and transaction‑risk pipelines that mirror global best practice: machine‑learning models that combine user‑behaviour signals, transaction history, device/IP telemetry and network analysis to assign a real‑time risk score, then trigger graduated responses from frictionless approval to OTP or manual review (fraud risk scoring overview, transaction risk‑scoring implementation guide).

Operationally, low‑latency feature stores and stream processors let those scores execute in milliseconds so legitimate payments - salary credits, microloan disbursements, social transfers - clear without delay while high‑risk flows are quarantined for investigation; Redis‑based pipelines are a proven pattern for serving online features and Bloom‑filter checks in real time (Redis for transaction risk scoring).

The payoff for Indonesia is practical: fewer false positives that lock out small merchants, stronger AML evidence trails for regulators, and analytics that surface underserved segments so financial‑inclusion strategies become data‑driven rather than guesswork.

A Badung Virtual Tourist Assistant - aligned with Sahabat‑AI efforts - would give Bali's crowded tourist corridors a multilingual, always‑on front desk that answers questions and helps with bookings around the clock, the same AI‑powered chatbot pattern already used by villa websites to assist guests 24/7 (AI-Powered Chatbots for Bali Villas); supporting English, Mandarin, Japanese and Russian lets the assistant meet visitors the moment they land, reduce pressure on local tourist offices, and nudge tourists toward safer choices or quieter neighborhoods without adding staff.

Framed as a public‑service tool, the assistant could also echo proven government pilots - prioritizing clear, auditable responses and local language coverage - as outlined in practical guides for deploying AI in Indonesia's public sector (Complete Guide to Using AI in Government in Indonesia), turning a late‑night booking query into a seamless, trustworthy interaction.

KemenpanRB's Policy Feedback Synthesizer could turn noisy, real‑time citizen chatter into actionable policy insights by applying the same Twitter‑based sentiment and aspect classification pipeline researchers used to compute Public Service Quality (PSQ) scores - automated models that convert polarity into a 1–5 PSQ scale, flag recurring complaints by aspect (time, software/apps, officers), and surface district‑level priorities for fast operational fixes; a Jakarta pilot using this method produced an overall PSQ of 1.70 (category D: Poor), showing how social feeds can spotlight service failures quickly (Evaluating public service quality through Twitter analysis).

Practical deployment would pair those classifiers (sentiment accuracy often 80–94% while aspect classification can sit near 40–47%) with bias‑correction and sampling weights, bot checks (the study found ~89% human‑like accounts) and audit logs so feedback informs policy reviews rather than drives reactionary edits; integrated dashboards could triage recurring software/app complaints to IT, route time‑related issues to process redesign teams, and feed procurement savings or e‑catalogue pilots into measurable targets (Complete guide to using AI in government in Indonesia (2025)).

The result is not just data but a tunable civic mirror - one that shows where services are failing in plain numbers and points to the operational fix most likely to reduce the next wave of complaints.

Score Range	Category	Meaning
4.51–5.00	A	Excellent
4.01–4.50	A‑	Very Good
3.51–4.00	B	Good
3.01–3.50	B‑	Good (With Notes)
2.51–3.00	C	Fair
2.01–2.50	C‑	Fair (With Notes)
1.51–2.00	D	Poor
1.01–1.50	E	Very Poor

@ccdukcapil @dukcapiljakarta mau ngurus NIK sm KK kok gak responsive bgt yah udh di email dr kmrn dan ditelfon, bahkan udah di WA dr kemarin.. tolong mba/mas dijawab.. saya mau daftar NPWP jd gak bisa krn no KK sm NIK gak sesuai.

Bangkit's industry‑university pilot offers a compact, government‑ready template for building AI capacity in Indonesia: an immersive, 900‑hour curriculum with three learning paths (Cloud, Android, Machine Learning) that pairs hands‑on capstones with employer connections so graduates are ready to work in public‑sector projects and reskilling programs (Bangkit Academy Google Cloud overview).

The program's scale and inclusivity - reported growth of 30×, more than 15,000 students trained, significant rural participation and strong industry partners - means ministries and regional governments can tap an existing talent pipeline for pilot deployments, teacher upskilling and curriculum co‑design rather than starting from scratch (Bangkit AI curriculum and national impact - Kompas).

Local case stories show the payoff: long, structured learning and capstone experience translate into real jobs and usable projects that public agencies can commission or adapt; for a human detail that sticks, graduates report completing hundreds of guided hours and collaborative capstones that mirror problems faced by puskesmas, transit agencies and municipal services (Bangkit student success story - Bangkit Academy blog).

Metric	Value
Curriculum length	900 hours / 18 weeks
Learning paths	Cloud, Android, Machine Learning (AI)
Reach	15,000+ students trained; cohorts of ~3,100 (Bangkit 2022)
Industry partners	Google, GoTo, Traveloka (and others)

“Keep manifesting, and let the universe speak.”

Introl and Telkom are well positioned to choreograph Indonesia's next wave of AI infrastructure by linking high‑density GPU orchestration with pragmatic site planning, resilient power design and hardened cyber‑response - a playbook that already appears in projects like BDx's renewable‑powered CGK4 AI campus (high power density, liquid‑cooling ready, Phase‑1 delivered in 75 days) and the Golden Digital Gateway in Batam (sub‑2ms links to Singapore, Phase‑2 adding ~20MW).

Coordinated orchestration means matching workloads to the right tier of site - central campuses for large‑scale training and island/edge sites for low‑latency inference - while baking in integrated power systems, modular cooling and identity‑centric defenses so outages or supply shocks don't cascade into citizen services.

The recent pauses by some hyperscalers in Jakarta underline the importance of flexible build plans and local partnerships; practical wins come from combining sovereign AI campuses, submarine‑cable connectivity and operator playbooks so compute is both performant and auditable for Indonesian public workloads (BDx CGK4 renewable-powered sovereign AI data center, Golden Digital Gateway Batam data center with sub-2ms Singapore connectivity).

Site / Feature	Key facts
BDx CGK4 (Jatiluhur)	Renewable‑powered AI campus; up to 120 kW/rack; liquid‑cooling ready; Phase‑1 deployed in 75 days; scalable toward 500MW
Golden Digital Gateway (Batam)	First Nongsa Digital Park facility; Phase‑2 ≈20MW expansion; 12 submarine cables enabling <2ms latency to Singapore; air‑cooled first phase

“Our ongoing partnership with Indosat is a transformative step in creating Indonesia's sovereign AI cloud while empowering the nation to be Southeast Asia's AI hub.”

Indonesia's playbook for turning AI ambition into public benefit is clear: lock policy and ethics into operational practice, fund fast pilots that prove value, and scale capacity across the civil service.

Start by adopting the new government evaluation mechanism and incident‑reporting approach to embed the draft AI Ethics Guidelines into everyday procurement and project review (Indonesia ethical AI evaluation mechanism), and use Presidential Regulation No.82 / SPBE indicators to prioritise interoperable data, secure infrastructure and monitoring across ministries (Indonesia digital government and SPBE roadmap).

Deliver quick wins via regulatory sandboxes and targeted pilots - for example, weather‑for‑agriculture and early‑warning health analytics - to build public trust and operational know‑how, while investing in workforce programs so thousands of civil servants can write better prompts, audit models and manage vendor risk (practical reskilling like Nucamp's 15‑week AI Essentials for Work accelerates this transition).

Pair pilots with continuous monitoring, clear self‑assessment pathways for developers, and a finance plan that blends state, private and international support so ethical AI moves from paper to routine, not just promise.

Program	Length	Cost (early bird)	Register
AI Essentials for Work	15 Weeks	$3,582	Register for AI Essentials for Work

IKN's Smart Governance strategy fosters transparency by making information accessible to the public, ultimately building trust and promoting a collaborative governance model.