Solution case studyAIOT-001

SIGNAL: paying vehicle tax and validating STNK without a counter, thanks to Police, Dukcapil, and Bapenda integration

AutomotiveVehicle registration services and regional vehicle tax
A mechanic using a vehicle diagnostic tool

Solution case studies are scenarios based on common industry problem patterns. They do not refer to any client and are not MarkasDev project portfolio items. Our client portfolio is published separately.

Industry context

Before the details: MarkasDev had no role in SIGNAL, and everything on this page is drawn from openly available documents and reporting. SIGNAL (Samsat Digital Nasional / National Digital Samsat) is a digital service developed by the Indonesian National Police Traffic Corps (Korlantas Polri) for annual vehicle registration certificate (STNK) validation, Motor Vehicle Tax (PKB) payment, and the compulsory road traffic accident fund contribution (SWDKLLJ). We read the official samsatdigital.id tutorial, Korlantas Polri articles, and reporting by ANTARA, detikOto, CNN Indonesia, Media Indonesia, JPNN, Pantau, and Tempo that quotes Korlantas officials. All sources are listed at the end.

SIGNAL was built as an improvement on Samsat Online Nasional (Samolnas). According to Police Grand Commissioner Taslim Chairuddin, then Kasubdit STNK at Korlantas (quoted by ANTARA, 20 August 2021, and detikOto, 21 August 2021), IT-based registration and identification services had been planned since 2014. Piloting began in June 2021 (available on the Play Store from 21 June 2021 in 15 Stage-I provinces). A 28 June 2021 launch planned under the National Police Chief’s first-100-days programme was delayed because of the pandemic/PPKM policy, then reframed as an Independence Day gift in August 2021. CNN Indonesia (25 August 2021) wrote that the official launch was scheduled around Traffic Police Day on 22 September 2021.

For operations leaders in fleet companies, local government/Bapenda offices, regional SOEs, insurers, logistics, and public services, SIGNAL is a concrete example of cross-agency integration: Police vehicle data, civil-registration identity data, provincial tax assessments, Jasa Raharja’s SWDKLLJ, settlement banks, and payment channels—with biometric AI checks so remote service still supervises vehicle ownership.

Problem

Before SIGNAL, people who needed annual STNK validation and to pay PKB plus SWDKLLJ usually had to visit a Samsat office, a mobile Samsat unit, a mall counter, or a drive-thru. The real cost was more than queue time. There was transport cost, exposure to brokers, and operational friction at counters. The official SIGNAL tutorial (samsatdigital.id) and a Korlantas article (7 May 2026) stress the goal of service “anywhere, anytime” without abandoning registration-and-identification oversight.

The technical problem runs deeper than “build a tax-payment app.” The required data sit in different institutions: the Police vehicle database (later described as the Regident Center / ERI), the civil registry held by the Directorate General of Dukcapil at the Ministry of Home Affairs, tax assessments at each provincial Bapenda, and SWDKLLJ / e-KD determination by Jasa Raharja. Without integration, remote service is exposed to misuse—vehicles already transferred, mismatched identity, or unsynced tax bills.

During the pilot, Korlantas itself acknowledged adoption friction. As of 13 August 2021, according to Taslim (ANTARA): 36,531 downloads, 18,860 vehicles registered in the app, and 5,131 transactions completed through payment (PKB value Rp6.9 billion and SWDKLLJ Rp403 million). CNN Indonesia (25 August 2021) cites 13,729 “failures” out of 18,860 (about 72.3 percent) when comparing registered vehicles with successful payments—then explains that many failures came from user procedure errors, tax not yet due, five-year STNK renewal cases, or provinces outside the 15 already connected to Bapenda. e-KTP verification success was reported at 85 percent, vehicle add at 78 percent, and validation transactions at 67 percent.

So the problem was not merely mobile UI. It was cross-agency data architecture, remote identity verification, province-by-province onboarding, and user education—exactly at the intersection of custom software, API integration, and supervised AI.

Constraints

  • Had to keep supervising vehicle ownership (regident), not only speed up payment (samsatdigital.id SIGNAL tutorial)
  • Had to integrate at least three national data sources: Police vehicle records, Dukcapil civil data, and provincial Bapenda tax systems (Korlantas 7 May 2026 and samsatdigital.id)
  • Remote identity checks use AI face matching against the e-KTP photo, plus SMS OTP and email verification (Korlantas, samsatdigital.id, detikOto)
  • Initially limited to privately owned vehicles (not legal entities). Signal Corporate is under development (detikOto and Tempo, October 2025)
  • PKB assessments follow each provincial Bapenda, so coverage depends on local integration (15 provinces in the 2021 pilot, 29 provinces per the Dirregident in December 2025)
  • Had to connect payment partners (Himbara and regional development banks), switcher reconciliation, Jasa Raharja, and optional physical TBPKP delivery (detikOto 21 August 2021, Espos/Solopos 5 September 2021)
  • Digital documents (e-TBPKP, e-validation, e-KD) must be legally usable. Espos cites e-validation authorisation by BSrE BSSN
  • Evidence limit: national build cost, aggregate uptime SLAs, post-2021 successful-transaction breakdowns by province, and the exact face-matching model are not fully published in the sources we opened

Approach

Diagnose

This section is our analysis of how a Diagnose phase shows up in the SIGNAL case. It is not a claim that MarkasDev delivered the work.

Diagnose starts from recognising Samolnas’ shortcomings and the inevitability of digital public registration services. Taslim (detikOto, 21 August 2021) lists foundation steps: consolidate vehicle data into the Korlantas Regident Center, build a national registration service (ERI), then SIGNAL integrated with Dukcapil, AI face recognition (with the National Police ICT Division), Bapenda, Jasa Raharja, regional development banks as settlement banks, payment partners, switcher reconciliation, and document-delivery partners.

In other words, Diagnose here is not “which app features look cool.” It is: which systems are sources of truth, which agencies must be connected, and which risk controls replace a physical counter visit (face matching + national ID number + OTP).

Diagnose questions for your organisation: which processes still force citizens or customers to show up only because backends are not integrated? Which identity can be verified remotely with acceptable risk? Which provinces/units must be onboarded first for value to appear? See our Diagnose approach.

Deliver

This is also analysis. Deliver, in the SIGNAL story, looks like a mobile platform hung on cross-agency integration contracts—not a big-bang replacement database for every party.

The official tutorial describes the user flow: registration (national ID number, e-KTP name, email, phone, password, e-KTP photo, biometric selfie, SMS OTP, email link), add vehicle (plate number + last five chassis digits, or a same-family-card scheme), display the payment assessment (PKB + SWDKLLJ), optional physical TBPKP shipping, generate a payment code, pay via partner banks, then receive e-TBPKP, e-validation, and e-KD in the app.

Early coverage (2021) targeted 15 provinces. Lists cited by ANTARA and Espos include DKI Jakarta, Banten, Java provinces, Bali, West Nusa Tenggara, and several Sumatra and Sulawesi provinces (early reports differ slightly on a few names—we do not force those lists into one invented roster). Payments involved state-owned Himbara banks (Mandiri, BNI, BRI, BTN) and regional development banks, plus optional delivery via Pos Indonesia (Espos, 5 September 2021).

Reasonable acceptance criteria for a similar programme (our analysis): identity verified without a counter visit, tax amounts matching official Bapenda channels, digital proofs that officers can check, and a clear path when verification fails. See also system & API integration, AI implementation, mobile applications, and Deliver.

Operate

This is also analysis. Operate, for SIGNAL, is the long work after go-live: expanding provinces, socialising the service, reducing verification friction, and adding corporate features.

As of October 2025, Korlantas Chief Inspector General Agus Suryonugroho said Signal had been downloaded by about 13 million users and still judged uptake “not yet large,” so revitalisation and outreach were needed (detikOto, 21 October 2025, quoting the press conference; Pantau and BeritaNasional cite the same download/user figure). Tempo (20 October 2025) reported the same briefing as “about 13 million transactions”—the user-versus-transaction wording difference across outlets should be noted. We do not “correct” figures beyond what each source wrote.

In December 2025, Dirregident Brigadier General Wibowo said SIGNAL coverage had reached 29 provinces, listing areas across Java–Bali–NTB, Sumatra, Kalimantan, Sulawesi, Maluku, Papua, and West Papua, with other regions still integrating (Media Indonesia, 2 December 2025, and JPNN, 3 December 2025).

October 2025 also pointed to Signal Corporate so legal entities with large fleets can pay collectively, with pilot work in Banten, West Java, and DKI Jakarta (Tempo). A Korlantas article on 7 May 2026 still stresses AI face matching, the three national databases, digital document issuance, and optional physical delivery.

In an Operate frame, the lesson for organisations is: a digital public service is not finished on launch day. Impact depends on regional onboarding, user education, monitoring verification failures, and expanding scenarios (personal → corporate) without weakening identity controls.

High-level architecture

This flow is our simplification of samsatdigital.id, the Korlantas article of 7 May 2026, and Taslim’s explanation in detikOto on 21 August 2021—not an architecture we audited. The AI named in sources is face matching/face recognition (including National Police ICT Division involvement in 2021 briefings), not a chatbot or tax-forecasting model. Inter-agency API protocols, national capacity, and the exact AI model are not fully published in the sources we opened. Service reach depends on each provincial Bapenda integration status. The figures of 15 provinces (2021) and 29 provinces (December 2025) come from Korlantas officials via media we opened.

Typical outcomes

The figures below come from sources we opened. We do not round beyond what those sources wrote.

Pilot through 13 August 2021 (ANTARA, quoting Police Grand Commissioner Taslim Chairuddin): 36,531 downloads, 18,860 vehicles registered in the app, 5,131 transactions completed through payment, PKB value Rp6.9 billion, SWDKLLJ value Rp403 million. Stage-I pilot in 15 provinces. Play Store availability from 21 June 2021.

Early process quality (CNN Indonesia, 25 August 2021, quoting Taslim): e-KTP verification success 85 percent, vehicle add 78 percent, validation transactions 67 percent. Comparing registered vehicles with successful payments yields a 13,729 gap (about 72.3 percent), largely explained as user procedure errors or rule mismatches (tax not due, five-year STNK, unconnected province)—not simply “the system was down.”

Adoption by October 2025 (detikOto, 21 October 2025, quoting Korlantas Chief Inspector General Agus Suryonugroho): Signal had been downloaded by about 13 million users. Pantau and BeritaNasional cite the same download/user figure from the revitalisation briefing. Tempo (20 October 2025) wrote “about 13 million transactions” for the same briefing—we report that wording difference as published.

Coverage by December 2025 (Media Indonesia 2 December 2025 and JPNN 3 December 2025, quoting Dirregident Brigadier General Wibowo): SIGNAL covers 29 provinces, with a long list of integrated regions and others still joining.

Product direction (Tempo and detikOto, October 2025): Signal Corporate for legal-entity vehicles, with pilots in Banten, West Java, and DKI Jakarta. The personal service remains focused on PKB, annual STNK validation, and SWDKLLJ.

Not found in open sources we opened: national build cost, aggregate uptime SLAs, detailed annual successful-transaction counts after 2021 on the official site, and face-matching model accuracy metrics.

Typical or expected outcomes, not claims about MarkasDev project results.

Further analysis: why this case matters for organisations in Indonesia

If you are mapping services that still force physical attendance only because backends are not connected, start with Diagnose, build integration through Deliver, and sustain adoption through Operate.

SIGNAL matters beyond STNK errands. The same pattern appears when hospitals connect EMR systems to patient identity, regional SOEs collect levies digitally, fleet companies chase vehicle-tax compliance, or financial institutions verify customers remotely. The core is many source systems, one user experience, and risk controls that must not disappear just because the counter became a phone screen.

Diagnose/Deliver/Operate lessons from this case:

  • Diagnose: map source agencies/systems, identity-fraud risk, and which provinces/units must onboard first.
  • Deliver: build the app on API contracts, face-matching AI with a clear failure path, and auditable digital proofs.
  • Operate: monitor downloads versus successful transactions, improve user education, expand regions, then add corporate personas.

Sources

  1. National Digital Samsat — official SIGNAL tutorial: https://samsatdigital.id/tutorial.php
  2. Korlantas Polri, 7 May 2026 — paying vehicle tax via SIGNAL: https://korlantas.polri.go.id/2026-05-07-lewat-aplikasi-signal-bayar-pajak-kendaraan-kini-lebih-mudah/
  3. ANTARA, 20 August 2021 — 36,531 downloads of National Digital Samsat: https://www.antaranews.com/berita/2341130/36531-pengguna-mengunduh-aplikasi-samsat-digital-nasional
  4. detikOto, 21 August 2021 — how Police improved Digital Samsat (Taslim integration architecture): https://oto.detik.com/berita/d-5690594/cara-polisi-sempurnakan-samsat-digital-demi-masyarakat-bayar-pajak-di-mana-aja
  5. CNN Indonesia, 25 August 2021 — Signal benefits, pilot figures, success rates: https://www.cnnindonesia.com/otomotif/20210824111142-579-684489/manfaat-dan-cara-pakai-aplikasi-signal-rilis-22-september
  6. Espos/Solopos, 5 September 2021 — how to use National Digital Samsat, 15 provinces, Himbara/BPD/Pos: https://otomotif.espos.id/cara-gunakan-samsat-digital-nasional-terintegrasi-15-provinsi-1151744
  7. detikOto, 21 October 2025 — Korlantas chief: about 13 million Signal downloads: https://oto.detik.com/berita/d-8170953/kakorlantas-bayar-pajak-kendaraan-semudah-beli-pulsa
  8. Pantau, 20 October 2025 — Signal/Sinar revitalisation, ~13 million downloads: https://www.pantau.com/nasional/299345/korlantas-polri-revitalisasi-layanan-digital-permudah-akses-pajak-dan-sim-lewat-aplikasi-signal-dan-sinar
  9. Tempo, 20 October 2025 — same briefing as ~13 million transactions, Signal Corporate: https://www.tempo.co/hukum/digitalisasi-pelayanan-publik-korlantas-targetkan-bayar-pajak-kendaraan-semudah-beli-pulsa-2081460
  10. Media Indonesia, 2 December 2025 — Signal reaches 29 provinces (Brig. Gen. Wibowo): https://mediaindonesia.com/politik-dan-hukum/836245/perkuat-quick-wins-korlantas-aplikasi-signal-kini-jangkau-29-provinsi
  11. JPNN, 3 December 2025 — SIGNAL strengthening and 29 provinces: https://m.jpnn.com/news/ditregident-korlantas-perkuat-inovasi-layanan-signal-sebagai-implementasi-quick-wins
  12. Korlantas Polri, 22 October 2025 — optimising SIGNAL, SINAR, e-BPKB: https://korlantas.polri.go.id/kakorlantas-dorong-optimalisasi-signal-sinar-dan-e-bpkb-untuk-layanan-cepat-dan-transparan/

Accessed on 13 October 2026 (WIB).

Common questions

Did MarkasDev build or operate SIGNAL?
No. SIGNAL is a Korlantas Polri service with stakeholders including Dukcapil, Bapenda, Jasa Raharja, and partner banks. This page summarises accessible documents and reporting, not a MarkasDev client or portfolio item.
When was SIGNAL piloted and launched?
Piloting from June 2021 (Play Store on 21 June 2021 in 15 Stage-I provinces per ANTARA). The official launch window was August–September 2021 after a delay from the planned 28 June date because of the pandemic (ANTARA, CNN Indonesia).
How many SIGNAL users are verified in sources we opened?
Early phase (13 August 2021): 36,531 downloads (ANTARA). By October 2025: about 13 million downloads/users according to the Korlantas chief as quoted by detikOto, Pantau, and BeritaNasional. Tempo wrote “about 13 million transactions” for the same briefing.
Does SIGNAL use AI?
Yes. Official sources (samsatdigital.id and Korlantas 7 May 2026) describe face matching/AI to compare a selfie with the e-KTP photo. detikOto (2021) cites AI face recognition with the National Police ICT Division.
What can non-Samsat organisations learn from this case?
Map source-of-truth systems first, design remote verification that replaces physical controls, measure failures with clear categories, then expand onboarding unit by unit (province/branch) before adding new personas such as corporate accounts.