Top 10 AI Prompts and Use Cases and in the Financial Services Industry in Indonesia

Indonesia's financial services scene is at an inflection point: with internet penetration near 79% and more than 180 million smartphone users scattered across 17,504 islands, AI is already improving fraud detection, alternative credit scoring and hyper‑personalized customer journeys while regulators push for safer, explainable systems - a careful balance the World Economic Forum maps in its review of AI-driven inclusion (World Economic Forum review of AI-driven inclusion in Indonesia).

From GenAI-powered underwriting to agentic onboarding that can verify ID and screen AML risks, the real question is who builds and operates these models responsibly; practical workplace skills - prompt design, model testing and human-in-the-loop workflows - are essential.

For teams wanting hands-on prompt training and applied AI skills tailored to business use, Nucamp's AI Essentials for Work bootcamp offers a 15‑week path to put those capabilities into practice and help turn promising pilots into safe, scalable services (AI Essentials for Work registration).

Attribute	Details
Program	AI Essentials for Work
Length	15 Weeks
Courses included	AI at Work: Foundations; Writing AI Prompts; Job Based Practical AI Skills
Early bird cost	$3,582
Registration	AI Essentials for Work registration

Methodology: to identify the Top 10 AI prompts and use cases for Indonesia's financial services sector, sources were triangulated across recent industry and policy reports and scored against three pragmatic lenses - adoption momentum, inclusion impact, and regulatory fit.

Evidence came from the World Economic Forum's review of AI-driven inclusion and the Indonesia Payment System Blueprint, Stanford HAI's 2025 AI Index for technical and safety benchmarks, and AWS's August 2025 research on adoption, revenue and cost impacts; each use case was prioritized when it demonstrated clear business value (AWS's revenue/cost metrics), measurable reach (WEF's internet, smartphone and financial‑inclusion stats) and alignment with central bank / OJK frameworks.

The resulting prompts favour fraud reduction, alternative credit scoring, conversational multilingual support and interoperable payment flows that can scale across Indonesia's islands and diverse user base.

Criteria	Evidence / Source
Adoption momentum	AWS research on AI adoption in Indonesia (August 2025)
Inclusion & reach	World Economic Forum analysis of AI-driven inclusion in Indonesia (February 2025)
Performance & safety	Stanford HAI 2025 AI Index report on AI performance and safety

“It is an interesting phenomenon we are seeing with AI adoption coming out of the study results in Indonesia. While 28% of businesses reported they have adopted AI, most of the deployments remain basic despite the rapid adoption of the technology over the past year. Larger enterprises are also at risk of being left behind by the nimbler, faster-paced startups. The resulting ‘two-tier' economy could have lasting implications on a country's future economic development. Celebrating AI adoption numbers alone masks the deeper challenges many businesses face across Indonesia.” - Nick Bonstow, Director at Strand Partners

DANA's approach to fraud and scam detection shows how Indonesian fintechs can move from rule‑based blocking to adaptive, real‑time defence: by continuously updating risk parameters and typology analysis, using Smart Friction to alert users at the moment they try to send funds to flagged accounts, and sharing data across agencies to close ecosystem gaps.

The results are striking - an 80% drop in gambling‑linked transactions, a digital patrol that has flagged over 39,000 websites and social accounts, and hundreds of thousands of suspicious users reported to Kominfo - all while working with Bank Indonesia and the Financial Intelligence Unit to align enforcement with law and AML expectations (see the coverage of DANA's program).

For institutions scaling beyond pilot projects, mature real‑time engines and link‑analysis tooling - the kind described in enterprise solutions like Eastnets' PaymentGuard - offer multi‑channel monitoring, custom risk scoring and lower false positives, which translates into fewer losses and faster investigations; the vivid takeaway is simple: stop the bad flow before it reaches a user's balance, not after.

“We are no longer reactive; we are predictive,” Vince Iswara, CEO and co‑founder of DANA operator PT Espay Debit Indonesia Koe.

Indonesia's persistent MSME credit gap - just 12.7% of MSMEs borrowed from banks or other sources in 2021 - makes innovative credit scoring (ICS) a strategic opportunity for incumbents like Bank Rakyat Indonesia (BRI): machine learning models that ingest alternative data (mobile footprints, app presence, social signals and bill payment histories) can reveal repayment signals traditional credit files miss, so that

“a phone's app list”

becomes part of a borrower dossier rather than a mystery; this practical pivot is laid out in evidence-based insights on using alternative data and AI for inclusion (Using Alternative Data and AI to Expand Financial Inclusion - Poverty Action Lab briefing).

Implementation research also shows ICS can raise scoring accuracy and process efficiency - but not without guardrails: data integration hurdles, consent and privacy under the Personal Data Protection Law, OJK's supportive stance (OJK Reg No.3/2024), and the risk of gender or sampling bias mean models should be paired with human review and validation frameworks described in the IJFS study on ICS implementation (Innovative Credit Scoring Implementation - IJFS article).

The vivid takeaway for BRI is clear - responsibly deployed ICS can open lending to previously invisible MSMEs, provided rigorous testing, explainability and data‑sharing standards move at the same pace as model accuracy.

Source	Key point
Using Alternative Data and AI to Expand Financial Inclusion - Poverty Action Lab (Mar 2024)	Alternative data and AI can expand MSME credit access; highlights privacy, bias and integration challenges.
Innovative Credit Scoring Implementation - IJFS (Aug 2025)	ICS improves process efficiency and scoring accuracy; recommends validation frameworks and ethical oversight.

DANA is turning its wide merchant footprint and device‑level signals into personalized financial advice and product recommendations that fit Indonesia's patchwork of users: by combining Ant Financial‑backed analytics and lightweight AI/ML with risk engines that use device data and transaction characteristics, the wallet can surface timely offers - from small eMAS gold purchases (as little as 0.01 gram, under IDR10,000) to bill‑payment reminders and SME payment tools - so rural vendors and domestic workers see actionable options when they need them most.

This blend of inclusion and productization relies on simple, local UX patterns (QR codes, low‑bandwidth flows) and partnership programs that teach use-cases like payroll and remittances; see the deep dive on DANA's rural strategy and tech choices for context and the collaboration that brought digital payroll to domestic workers.

With strong security measures and protections baked in, the vivid takeaway is practical: micro‑offers - tiny gold, a targeted savings prompt, or a tailored merchant discount - can turn a single tap into sustained financial engagement across Indonesia's islands (DANA Payments guide, Women's World Banking collaboration).

Attribute	Value
Payment Method Category	eWallet
Country	Indonesia
Presentment Currency	IDR
Minimum Transaction	10,000 IDR
Settlement Timeframe	2–3 days

“As an Indonesia's digital wallet that focuses on pursuing an inclusive digital economy through best‑in‑class technology, DANA strengthens the effort by joining forces with Women's World Banking to encourage women, especially domestic workers, to be more capable in digital financial. We would like to instill that digital wallet can be used easily by anyone – including domestic workers – to help their families safely and conveniently.” - Monita Moerdani, Chief Marketing Officer, DANA

Complementing wallet-led personalization, GoTo's Dira introduces conversational AI in Bahasa Indonesia to make everyday finance truly voice‑first: a lightweight assistant embedded in the GoPay app that lets users tap the microphone icon and, for example, reach the BPJS insurance‑payment page in a single command instead of multiple taps and scrolls.

Designed as part of GoTo AI's roadmap to improve convenience, safety and AI capability across the ecosystem, Dira (short for Dikte Suara) is rolling out to a limited user group now with plans to expand to Gojek, and its low‑footprint integration means even low‑capacity phones can access voice navigation at no extra cost - an important inclusion lever for Indonesia's large unbanked population.

Security remains layered: transactions still require PIN or biometric confirmation, while the feature's emphasis on Bahasa and simple flows aims to lower friction for older users, informal workers and those with limited literacy.

For product teams, the key lesson is practical: localized voice UX can shrink multi‑step journeys into a single, memorable interaction that broadens access and speeds adoption (GoTo launches Dira AI voice assistant - official press release, Fintech News Singapore coverage of GoTo's Dira AI voice assistant rollout).

“Dira by GoTo AI delivers a localized voice assistance solution that not only benefits our users, but also paves the way for more AI features across the GoTo ecosystem in the future.” - Hans Patuwo, GoTo Chief Operating Officer

KYC, identity verification and AML monitoring in Indonesia are being recast as an integrated, OJK‑guided stack where identity proofing, transaction surveillance and AI governance must move in step: OJK's Artificial Intelligence Governance for Indonesian Banks sets the baseline for explainability, risk management and human‑in‑the‑loop controls for models used in onboarding and screening, while POJK 8/2023 expands AML obligations across fintechs and requires firms to report risk assessments and tighten CDD procedures (OJK Artificial Intelligence Governance for Indonesian Banks, POJK 8/2023 AML‑CFT‑CPF regulation).

Practical implementations must also meet operational timings and thresholds already familiar to compliance teams - suspicious transaction reports are expected within three working days and cash transaction reporting kicks in at IDR 500 million - so any Sahabat‑AI style solution that automates eKYC and transaction scoring must bake in fast, auditable case management and data retention rules from the outset (AML and KYC compliance in Indonesia - Sumsub guide).

The vivid takeaway: AI can speed identity checks, but regulators demand the audit trail - explainable decisions, human review and timely STR filing - so resilience is measured not just by automation speed but by compliance‑grade traceability.

Requirement	Detail
AI governance baseline	OJK Artificial Intelligence Governance for Indonesian Banks - explains explainability, risk management and human oversight
AML regulation	POJK 8/2023 - expands AML/CFT/CPF scope and reporting obligations for financial service providers
STR timeline	Report suspicious transactions within 3 working days
Reporting threshold	Cash transactions ≥ IDR 500,000,000 must be reported

PasarPolis illustrates how insurance claims in Indonesia can move from slow, paper‑heavy workflows to near‑instant, scalable automation by embedding microinsurance into everyday apps and leaning on cloud‑native tooling: with Google Cloud's GKE, BigQuery and Vision/Dialogs AI the platform handles terabyte‑scale risk analytics and can absorb sudden three‑to‑four‑fold demand spikes while freeing engineers from half their DevOps chores (PasarPolis Google Cloud case study on scaling insurance with GKE, BigQuery, and AI).

The result for users is tangible - tiny, contextual policies sold at the point of need and one‑tap claims for simple events - and for partners it's enterprise throughput (PasarPolis has scaled to tens of millions of policies monthly and integrates directly into ecosystems like GoTo/Gojek to issue policies at high velocity) as documented in coverage of the GoTo partnership and operational benchmarks (PasarPolis GoTo/Gojek partnership report expanding insurance access in Indonesia).

Strategic backers and development partners such as the IFC have also supported platform expansion to reach underserved Indonesians, underscoring a core insight: cloud‑first automation plus embedded distribution can turn claims from a barrier into the moment customers actually experience insurance protection (IFC press release: IFC and PasarPolis partnership to boost microinsurance).

Metric	Value / Source
Monthly policies	70 million (Google Cloud case study)
Total policies issued (reported)	775+ million (IFC press release)
PasarPolis Mitra agents	10,000 (LeapFrog case study)
Key cloud tools	GKE, BigQuery, Dialogflow, Vision AI, Pub/Sub, Cloud Storage
Series B funding	US$54M (investors incl. Leapfrog, Xiaomi, GoVentures)

“With Google Kubernetes Engine, a sudden three‑ or fourfold spike in demand is automatically absorbed without our team having to do a thing, and the system scales down just as quickly.” - Nishant Kumar, CTO, PasarPolis

Bank Indonesia's 2030 Payment System Blueprint is turning payments orchestration and routing optimization into a nationwide platform play: by standardizing APIs, introducing a Payment ID as a granular transaction “fingerprint,” and building BI‑Payment Clear and BI‑Payment Info as clearing and data‑as‑a‑service layers, the Blueprint aims to route flows with lower latency, clearer liability and stronger risk controls while matching access and supervision to an operator's size and interconnectedness.

That means orchestration engines must combine real‑time routing, failover to central rails, and intelligent cost/pricing logic so payments take the safest, cheapest path across islands; add cross‑border QRIS expansion and the Digital Rupiah pilot, and the same routing fabric will need to support retail, wholesale and CBDC‑anchored flows.

For product and compliance teams the takeaway is vivid: once Payment ID and BI‑managed services are live, a single misrouted payment will be traceable end‑to‑end - turning routing optimization from cost play into a compliance and consumer‑protection imperative (see the Bank Indonesia 2030 Payment System Blueprint and the BSPI five‑initiative overview for implementation context).

BSPI Initiative	Purpose
Bank Indonesia 2030 Payment System Blueprint – Infrastructure initiative (Baker McKenzie analysis)	Modernize retail/wholesale rails, BI‑Payment Clear, Payment ID and BI‑Payment Info.
Industry	Align access, licensing and supervision to participant risk and market contribution.
Innovation	Promote safe payments innovation via a Digital Innovation Center and consumer protections.
International	Expand cross‑border connectivity (QRIS cooperation, interoperable rails).
Digital Rupiah	Advance wholesale CBDC experiments and use cases for market infrastructure.

Sahabat‑AI is emerging as a practical lever for localized financial education and inclusion in Indonesia: this homegrown, open‑source LLM - now a 70‑billion‑parameter model - speaks Bahasa Indonesia plus Javanese, Sundanese, Balinese and Bataknese, and is available as a multilingual chat service on Sahabat‑AI official site (multilingual chat service) and via the GoPay “Popular Services” tab, lowering linguistic barriers for millions of users who need simple, context‑aware explanations of bills, savings, microinsurance and payment flows; because the model and its GPU Merdeka infrastructure keep data and compute inside Indonesia, banks and fintechs can build compliant, culturally contextual tutors and voice assistants that run on accessible hardware (the optimized model can operate on just two H100 GPUs), making financial guidance feel local rather than foreign and turning a one‑line explanation into a teachable moment for a vendor on a remote island.

Learn more on the Sahabat‑AI project page (official) and Light Reading's coverage of the multilingual rollout.

Attribute	Detail / Source
Model size	70 billion parameters (Sahabat‑AI official site (model details))
Languages	Bahasa Indonesia, Javanese, Sundanese, Balinese, Bataknese (Telecom Review Asia article on Sahabat‑AI multilingual support)
Availability	Chat service on sahabat-ai.com and GoPay app (Light Reading article on Sahabat‑AI multilingual chat service)
Infrastructure	GPU Merdeka, locally hosted, operates on two H100 GPUs (Indosat/GoTo disclosures)
Downloads	35,000+ for earlier models on Hugging Face (reported)

“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.” - Patrick Walujo, GoTo Group CEO

Regulatory reporting in Indonesia is moving from periodic checklists to near‑real‑time supervision, and OJK's recent toolkit makes the stakes crystal clear: POJK 12/2024 tightens anti‑fraud transparency with fast incident windows (as short as three business days for banks) and steep per‑day penalties (up to IDR 1.5 million/day for commercial banks), while ownership disclosures under POJK.4/2024 require prompt filing within five working days and will shorten further once an electronic system is live; banks and fintechs must therefore stitch together explainable model outputs, automated timelines and tamper‑proof audit trails to avoid fines and reputational damage.

Practical compliance now blends fraud detection with regulatory workflows - advanced threat detection and automated reporting help map alerts into STRs and regulator submissions - so security telemetry, case management and machine‑readable evidence become governance primitives rather than optional add‑ons (OJK POJK 12/2024 regulatory overview, POJK.4/2024 ownership reporting requirements in Indonesia, Exabeam guidance on automated detection and regulatory timelines).

The vivid takeaway: a single missed deadline can trigger daily fines and spiral into license risk, so automation plus auditable trails is the new baseline for any AI‑driven product in Indonesia's financial sector.

Regulation / Initiative	Key requirement	Deadline / Penalty
POJK 12/2024	Fraud reporting, enhanced anti‑fraud strategy and transparency	Fraud incident reports: 3 business days for banks; penalties up to IDR 1.5M/day
POJK.4/2024 (Ownership)	Report changes in voting rights and encumbrances	Submit within 5 working days; electronic system will shorten to 3 business days
Electronic reporting upgrade (KSEI)	Centralised e‑reporting system for ownership/encumbrance filings	KSEI to serve as provider once implemented (reduces electronic submission timelines)

GoTo and Tokopedia turn customer segmentation into a repeatable growth machine by marrying Tokopedia's Customer Data Platform - built on BigQuery and Bigtable with self‑service analytics and unified attributes - to GoTo's ecosystem signals so merchants, drivers and diverse consumer cohorts (urban and semi‑urban, digitally savvy or low‑bandwidth users) see precisely timed, localized offers; the CDP's single definition of attributes shortens a shopper's hunt through thousands of listings into a tiny, relevant shortlist and powers hyper‑local promotions and loyalty nudges that raise conversion and GMV. This data fabric helped fuel broad platform momentum - Group GTV rose to Rp116.5 trillion (20% YoY) while core GTV jumped 32% YoY - and underpins faster fintech scaling (outstanding loans grew 43% QoQ to Rp2.7 trillion), aided by cross‑sell channels like GoPay (20M+ downloads).

For product teams, the lesson is practical: invest in a governed CDP and segmentation stack to turn behavioral signals into product design, personalized promos and measurable GMV lift (see Tokopedia's CDP build and GoTo's results for implementation cues).

Metric	Value / Source
Group GTV (1Q24)	Rp116.5 trillion (GoTo Group Q1 2024 results press release)
Group core GTV	Rp54.6 trillion (32% YoY) (GoTo Group Q1 2024 results press release)
Tokopedia Monthly Active Users	100+ million (CDP context) (Tokopedia Customer Data Platform case study on Google Cloud)
GoPay downloads	20+ million (GoTo Group Q1 2024 results press release)
Outstanding consumer loans	Rp2.7 trillion (43% QoQ) (GoTo Group Q1 2024 results press release)

“We expect even faster growth for the rest of the year, while also remaining committed to our profitability goals.” - Patrick Walujo, GoTo Group CEO

Indonesia's AI moment is not theoretical - it's a practical test: across 17,504 islands and a market of over 280 million people with 180+ million smartphones, the tech and policy pieces must move together so innovation benefits everyone, not just a few urban hubs.

The right mix? pragmatic governance, explainable models that regulators can audit, and a worker‑centric push to reskill teams so underwriters, compliance officers and product managers can operationalize safe AI at scale.

Policymakers and industry are already laying rails - the World Economic Forum maps how AI is expanding inclusion while flagging data and bias risks (World Economic Forum: The rise of AI in Indonesia (2025)) - and national strategy plus private investment create a runway for both startups and incumbents.

For practitioners who need hands‑on prompt design, responsible testing and human‑in‑the‑loop workflows, a structured training path such as Nucamp's AI Essentials for Work helps teams turn pilots into compliant, customer‑safe products (Nucamp AI Essentials for Work bootcamp registration).

Indicator	Value
Population	Over 280 million
Smartphone users	Over 180 million
Internet penetration (2024)	79%
Financial inclusion index	~84%
Projected GMV by 2030	$200–360 billion

“whichever country ‘controls AI can potentially control the world'” - President Joko Widodo