AI to Prepare Students for the New TKA
Say a Grade 8 class has spent years hearing that the national assessment mostly checks reading and numeracy skill, not textbook content — and now that assumption is out of date. Indonesia's Tes Kemampuan Akademik (TKA), rolled out nationally by the Ministry of Primary and Secondary Education (Kemendikdasmen) for 2026, tests specific subject content directly.
At the SD/MI and SMP/MTs levels that means Bahasa Indonesia and Matematika; at the SMA/SMK level, English and two elective subjects join the list. AI tools can help students and teachers review that content and rehearse test-style questions, but they cannot substitute for a curriculum-anchored study plan.
Quick Answer: TKA is a new subject-based national assessment from Kemendikdasmen, replacing the older literacy/numeracy-only testing model with direct subject content review — Bahasa Indonesia and Matematika for SD/SMP, plus Bahasa Inggris and two elective subjects for SMA/SMK. AI is most useful for generating practice questions and content review material matched to those exact subjects; it cannot replace the official TKA blueprint or a teacher's judgment of what a specific class still needs to master.
This guide sits inside a broader look at how AI's usefulness shifts by country and assessment system — see AI in Education Around the World: A 2026 Regional Guide for the fuller picture across dozens of comparable contexts.
What the New TKA Actually Tests
TKA is Kemendikdasmen's replacement for the subject-content testing that Indonesia's education system had been missing since the Ujian Nasional (UN) was discontinued in favor of the literacy- and numeracy-focused Asesmen Nasional (AN). Its assessment framework for the SD and SMP levels was formalized through Peraturan Kepala Badan Standar, Kurikulum, dan Asesmen Pendidikan Kemendikdasmen No. 047/H/AN/2025.
From UN to AN to TKA: Why the Format Keeps Changing
Indonesia discontinued the Ujian Nasional in 2021 after years of criticism that a single high-stakes exam narrowed teaching toward test tricks rather than genuine subject mastery. Its replacement, Asesmen Nasional, deliberately dropped subject-content questions in favor of measuring literacy and numeracy skill more broadly through the Asesmen Kompetensi Minimum (AKM).
That shift solved one problem and created another: schools gained a less narrowly test-driven measurement tool, but the system lost any standardized signal of how well students actually knew specific subject content. TKA is Kemendikdasmen's answer to that gap — a return to subject-based testing, but built as a validator alongside classroom grades rather than a single make-or-break exam the way UN once functioned.
UN tested subject content and was criticized for narrowing instruction toward the exam. AN/AKM tested general literacy and numeracy instead. TKA tries to keep subject content while avoiding a single make-or-break score.
For a teacher who has only taught under the AN/AKM model, this is a real shift in what "test prep" means. AKM-style preparation focused on general reading and reasoning skill; TKA preparation looks more like ordinary subject review — closer to how many teachers already prepare students for a unit test, just applied at a larger, standardized scale.
Subjects and Timing by Level
The subject list and schedule differ meaningfully by level, and a teacher preparing students needs to know which one applies:
| Level | Subjects Tested | Planned Window |
|---|---|---|
| SD/MI (elementary) | Bahasa Indonesia, Matematika | March–April 2026 |
| SMP/MTs (junior high) | Bahasa Indonesia, Matematika | March–April 2026 |
| SMA/MA/Paket C, SMK/MAK (senior high) | Bahasa Indonesia, Matematika, Bahasa Inggris (compulsory) + 2 elective subjects | 26 October–8 November 2026 (makeup: 16–29 November) |
Students receive a Sertifikat Hasil TKA (SHTKA) — a results certificate — rather than a simple pass/fail outcome, and results for the senior-high window are expected around 23 December 2026.
Why TKA Doesn't Replace UTBK for University Entrance
A common mix-up worth clearing up early: TKA and UTBK are separate systems serving separate purposes. TKA results feed the SNBP pathway (Seleksi Nasional Berdasarkan Prestasi, the merit/report-card-based university admission route), acting as an objective check on a student's report-card grades rather than replacing them. The report card remains the primary component.
The SNBT pathway (Seleksi Nasional Berdasarkan Tes) still runs on UTBK — the computer-based test covering Tes Potensi Skolastik (TPS) and literacy, unchanged by TKA's arrival. A student aiming at SNBT admission still needs a separate UTBK preparation track; TKA prep and UTBK prep are not interchangeable, even though both eventually matter for the same university application.
Building an AI-Assisted TKA Study Plan
The most useful starting point is the subject list itself, not a generic "help my child study for a test" prompt. TKA's compulsory subjects are fixed and known well in advance, which makes targeted content review realistic rather than guesswork.
Start From the Subject List
- Confirm the level — SD/MI, SMP/MTs, or SMA/SMK/MA — since the subject list and timing both change by level.
- Pull the current curriculum scope for Bahasa Indonesia and Matematika (and Bahasa Inggris plus electives at the SMA/SMK level) for the specific grade.
- Ask an AI tool to draft practice questions matched to that scope, not a generic "TKA practice test" the tool has no real blueprint for.
- Have the teacher or a subject-competent adult check every generated answer key before students rely on it.
- Track which topics still need review after each practice round, so later sessions target gaps instead of repeating covered ground.
A Study Cadence That Fits Around Regular Coursework
Picture a Grade 8 Bahasa Indonesia and Matematika teacher building short review sets into normal weekly lessons rather than adding a separate cram period close to the March–April window. A workable pattern: one AI-generated short practice set per subject per week, reviewed together in class, with weaker topics flagged for a slightly longer follow-up set the next week.
That steady cadence tends to hold up better than a single intense review push, since TKA content review is closer to normal subject mastery than to a one-off exam-format drill. Consistency across the term matters more than a last-minute sprint.
Differentiating Practice Sets Within One Classroom
A single SD or SMP classroom often spans a wide range of Bahasa Indonesia and Matematika readiness, especially in a year when TKA is new for students and teachers alike. AI's ability to generate a simpler and a more advanced version of the same topic from one prompt is a genuine time-saver here.
- Ask for the same topic at two difficulty levels in one request, rather than writing two sets from scratch
- Keep the underlying topic identical across both versions, so a whole-class review discussion still works for everyone together
- Save both versions once verified, so the next unit on a related topic starts from an existing base instead of a blank page
Subject-by-Subject: Where AI Prep Helps Most
AI's reliability for generating TKA-style practice content varies by subject, largely for the same reasons it varies across any content-generation task: some subjects have more objectively checkable answers than others.
| Subject | AI Reliability for Practice Questions | What Still Needs a Human Check |
|---|---|---|
| Matematika | High — arithmetic, geometry, and algebra problems are easy to verify | Alignment with the exact grade-level curriculum scope |
| Bahasa Indonesia | Moderate — grammar and comprehension items are usable drafts | Idiom, regional usage, and question phrasing that matches real TKA style |
| Bahasa Inggris (SMA/SMK) | Moderate-high — vocabulary and grammar items are reliable | Register and cultural context in reading passages |
| Elective subjects (Science/Social Science tracks) | Moderate — depends heavily on subject | Factual accuracy on specialized topics; always spot-check |
For math practice specifically, Best AI for Math Problems in 2026 (Benchmarked) compares how different AI tools handle problem generation and step-by-step accuracy — useful background before leaning on any single tool for Matematika review.
Where AI Prep Falls Short for TKA
Two limits matter enough that a teacher or parent should plan around them rather than discover them mid-preparation.
AI Has No Access to the Actual TKA Blueprint
No AI tool has been trained on Kemendikdasmen's specific item bank or exact question format for TKA, because that material isn't public in a form models can learn from. Any AI-generated "TKA practice question" is an educated approximation of subject content, not a released past paper. Treat it as review material for the underlying subject, not a simulation of the real test's exact difficulty or phrasing.
The Report-Card Half of the Equation Is Outside AI's Reach Entirely
Because TKA functions as a validator for report-card grades under the SNBP pathway, a student's actual classroom performance across the term still carries the real weight. AI-generated practice sets can reinforce subject content, but they cannot substitute for consistent classwork, homework completion, and the ongoing formative assessment a teacher already runs.
A strong TKA result and a weak report card send Kemendikdasmen's validation system a mixed signal. The two need to move together, not one instead of the other.
What This Means for SD and SMP Teachers Specifically
TKA now reaches every SD/MI and SMP/MTs classroom, not just exam-track secondary students preparing for university. That's a real shift for teachers who have never had to prepare younger students for a standardized subject test — and it raises a fairness question, since Indonesia's regions don't all have equal access to the technology a fast AI-assisted workflow assumes.
A New Kind of Preparation for Younger Students
For many SD teachers, TKA is the first standardized, subject-content assessment their students will face outside normal classroom testing. Keeping review low-stress matters as much as keeping it accurate.
- Frame practice sets as normal classwork, not a special "big test" activity, especially for younger SD students encountering this format for the first time
- Keep individual sessions short — 10 to 15 minutes fits a Grade 5 school day far better than a long test-prep block
- Talk through wrong answers as a class, since the goal is subject mastery, not just correcting an answer key
Connectivity Gaps Across Indonesia's Regions
World Bank research on Indonesia's education sector has repeatedly flagged an uneven digital divide, with device access and reliable connectivity lagging in eastern provinces relative to Java and Bali (World Bank, 2023). A generate-once, print-and-reuse workflow matters more in a lower-connectivity school than one assuming constant live access to an AI tool.
- Batch-generate a term's worth of practice sets during a period of reliable connectivity, rather than generating fresh material lesson by lesson.
- Print physical copies as a backup for schools where device access is shared or limited.
- Build a shared, verified question bank with colleagues so the same connectivity-dependent generation step doesn't repeat across every classroom in a school.
Indonesia's broader assessment context adds weight to this: the OECD's PISA results have consistently placed Indonesian students below the OECD average across reading, math, and science (OECD, 2023), which is part of why a steady, well-supported subject-review habit matters for reasons that go beyond TKA alone.
Tools and Technology for TKA Preparation
Matching the tool to the task keeps a TKA study plan from turning into a scattered mix of unrelated apps.
| Tool Type | Best For | Caution |
|---|---|---|
| General AI assistant (Gemini, ChatGPT, Claude) | Drafting subject-specific practice questions in Bahasa Indonesia | Always verify against the actual grade-level curriculum scope |
| EduGenius | Generating differentiated Matematika and Bahasa Indonesia worksheets and quizzes matched to a class profile, with answer keys included | Strongest for structured, curriculum-anchored practice sets, not a substitute for the official TKA item bank |
| SCERT/Kemendikdasmen curriculum documents | Highest accuracy for confirming exact grade-level scope | No AI generation — source material only |
| A subject teacher's review pass | Final accuracy and phrasing check on any AI-generated set | Non-negotiable before distributing to students |
Because it uses class-profile settings for grade level and subject, a teacher could use EduGenius to generate a short Matematika practice set and an accompanying answer key in minutes, then adjust the difficulty for students who need extra support ahead of the March–April window. Its multi-format export means the same set can go out as a printed worksheet or a digital file depending on what a specific classroom needs.
For entrance-exam contexts with a longer AI-benchmarking track record, AI for ECAT and Engineering Entry Tests shows how a more established, high-stakes test format handles AI-assisted preparation — a useful comparison point even though TKA's structure and purpose are different.
Pro Tips for AI-Assisted TKA Prep
- Name the level and subject explicitly in every prompt — "SMP Grade 8 Matematika" produces far more usable output than "TKA practice questions."
- Generate small batches and review immediately, rather than producing a large set and hoping most of it is correct.
- Keep a running bank of verified questions by topic, so a checked item from one term is ready to reuse the next.
- Pair AI-generated Matematika drills with genuinely checkable answer keys — math is the subject where verification is fastest and most reliable.
- Loop electives in only after the compulsory subjects are covered, since Kemendikdasmen's compulsory-subject list is fixed and known well ahead of the test window.
- Batch-generate review sets whenever connectivity allows, especially in regions where reliable internet access is inconsistent, and print copies as a backup.
What to Avoid
- Don't present AI-generated questions as official TKA past papers. No public item bank exists for models to have learned from, so any resemblance to the real test is approximate at best.
- Don't neglect the report-card half of preparation. TKA validates classroom performance under SNBP — it doesn't replace the need for strong, consistent coursework.
- Don't mix up TKA prep with UTBK/SNBT prep. They test different content, serve different admission pathways, and need separate study plans.
- Don't skip the human accuracy check on generated answer keys, especially for Bahasa Indonesia comprehension items where phrasing and regional usage matter.
- Don't frame TKA as a frightening new high-stakes exam in front of young SD students. Presenting it as normal subject review keeps anxiety low and keeps the focus on genuine learning rather than test panic.
Key Takeaways
- TKA is Kemendikdasmen's new subject-based national assessment, formalized for SD/SMP through Regulation No. 047/H/AN/2025, testing Bahasa Indonesia and Matematika at SD/SMP and adding Bahasa Inggris plus two electives at SMA/SMK.
- SD/SMP TKA is planned for March–April 2026; the SMA/SMK window runs 26 October–8 November 2026, with results expected around 23 December 2026.
- TKA results validate SNBP report-card admissions — they are not part of SNBT, which still relies on UTBK's TPS and literacy format.
- AI is most reliable for Matematika practice generation and moderately reliable for Bahasa Indonesia and Bahasa Inggris content, but every generated item needs a human accuracy check.
- No AI tool has access to the real TKA item bank, so generated practice questions are approximations of subject content, not simulated past papers.
- A steady weekly review cadence tends to outperform a last-minute cram push, since TKA content review overlaps heavily with normal subject mastery.
Frequently Asked Questions
What is the TKA in Indonesian education?
TKA (Tes Kemampuan Akademik) is a national academic assessment introduced by Kemendikdasmen, testing subject content directly — Bahasa Indonesia and Matematika at the SD/SMP levels, with Bahasa Inggris and two elective subjects added at the SMA/SMK level.
Does TKA replace the UTBK university entrance test?
No. TKA feeds the SNBP (merit-based) admission pathway as a validator of report-card grades, while SNBT (the test-based pathway) continues to rely on UTBK's TPS and literacy format — the two systems run in parallel, not as replacements for each other.
When is TKA held in 2026?
SD/MI and SMP/MTs students are scheduled for a March–April 2026 window, while SMA/MA/Paket C and SMK/MAK students sit the main test 26 October–8 November 2026, with a makeup window 16–29 November and results expected around 23 December 2026.
Can AI tools generate accurate TKA practice questions?
AI tools can generate subject-relevant practice questions for Matematika and Bahasa Indonesia reasonably well, since neither Kemendikdasmen's exact item bank nor question style is public. Treat any AI-generated set as general subject review, always checked by a teacher, rather than a simulated real test.
Does EduGenius offer TKA-specific practice tests?
EduGenius does not have access to Kemendikdasmen's official TKA item bank, since it isn't public. It can generate Matematika and Bahasa Indonesia worksheets and quizzes matched to a class profile's grade level, which a teacher can use as general subject-review material alongside official guidance.
Is TKA harder to prepare for in schools with limited internet access?
AI-assisted review works best with reliable connectivity, but a batch-generate-and-print workflow — preparing a term's worth of material during a period of good access — keeps preparation realistic for schools in Indonesia's lower-connectivity regions, an equity gap the World Bank has documented in the country's education system (World Bank, 2023).
Sources
- Kemendikdasmen (Kementerian Pendidikan Dasar dan Menengah) — Peraturan Kepala Badan Standar, Kurikulum, dan Asesmen Pendidikan No. 047/H/AN/2025, TKA assessment framework for SD and SMP levels.
- Badan Standar, Kurikulum, dan Asesmen Pendidikan (BSKAP) — TKA subject lists, scheduling, and SHTKA certificate details.
- SNPMB (Sekretariat Penerimaan Mahasiswa Baru) — SNBP and SNBT pathway requirements and UTBK format for the 2026 admission cycle.
- OECD — PISA 2022 results placing Indonesian students below the OECD average across reading, math, and science (published 2023).
- World Bank — research on Indonesia's education-sector digital divide and regional connectivity gaps (2023).
- ISTE (International Society for Technology in Education) — guidance on human review of AI-generated instructional and assessment content.