AI Lesson Plans Aligned to Asesmen Nasional

Asesmen Nasional (AN) doesn't grade what a student memorized this term — it samples reading and math reasoning skills built up over years, tested only at grade 5, 8, and 11. AI lesson planning aligns best when it strengthens those underlying literacy and numeracy habits every week, across every subject, rather than cramming a review sprint the month before testing season.

Quick Answer: Asesmen Nasional measures reading literacy and numeracy reasoning (AKM), classroom character, and school learning environment — not individual subject content — and only a sample of grade 5, 8, and 11 students sit it each year. AI lesson plans align best by building AKM-style reasoning (multi-text comprehension, applied number sense) into ordinary weekly lessons across all grades and subjects, since the skills tested are cumulative, not something a single test-prep unit can produce.

Indonesia's education ministry restructured in October 2024, splitting the former Kemendikbudristek into three bodies under Presidential Regulation No. 139/2024. School-level policy, including AN, now sits with the Ministry of Basic and Secondary Education (Kemendikdasmen), led by Minister Abdul Mu'ti. The assessment itself hasn't changed shape — it's still the same three-instrument model introduced in 2021 to replace the old Ujian Nasional (UN).

What Asesmen Nasional Actually Measures

AN photographs a school's overall quality rather than certifying any individual student. That single design choice — a diagnostic snapshot, not a graduation gate — is the thing most lesson-planning advice gets wrong by treating it like the old UN.

Three Instruments, One School-Level Picture

Every school takes part, but only a sample of students inside it does. The three instruments feed into a school's Rapor Pendidikan (Education Report Card), a public dashboard meant to guide improvement rather than rank students against each other.

Sample-Based by Design, Not Every Student, Every Year

Sampling caps are specific and small relative to a full cohort: up to 30 students at grade 5, and up to 45 each at grades 8 and 11. A school below that threshold tests roughly 85% of its eligible students instead.

• Grade 5 (SD/MI): 30 questions each for reading literacy and numeracy.
• Grade 8 (SMP/MTs) and grade 11 (SMA/MA/SMK): 36 questions each for the same two domains.
• Every school participates at the institutional level — the sampling applies to students, not schools.

Because any student in the eligible grade could be selected, and because AKM measures reasoning built up over years rather than a single term's content, a lesson plan that only "preps for AN" in the weeks beforehand is targeting the wrong unit of time.

Asesmen Nasional vs. the Old Ujian Nasional

Comparing the two side by side is the fastest way to see why an AN-aligned lesson plan looks nothing like old UN test-prep drilling.

The shift matters for planning: UN rewarded intensive content review in the final term before the exam. AN rewards steady reasoning practice across every term, because the sampled cohort is never announced in advance and the skills measured don't come from any single unit.

The AKM Framework: What "Literacy" and "Numeracy" Actually Mean Here

AKM doesn't test a subject's factual content the way an end-of-unit quiz does. It tests whether a student can locate, interpret, and evaluate information in an unfamiliar text, and apply number sense to a real situation — skills a student either has or doesn't, regardless of which chapter was covered last.

Reading Literacy: Multiple Text Types, Not Recall

AKM's reading passages mix fiction and informational texts, and questions typically move through three cognitive layers: finding explicit information, interpreting and integrating ideas across a passage, and evaluating or reflecting on content critically. A student who can only recall a textbook passage verbatim struggles here far more than one who reads unfamiliar material comfortably.

Numeracy: Content, Process, and Context Together

Numeracy questions generally draw on number, algebra, geometry and measurement, and data and uncertainty as content areas, applied through everyday contexts rather than as isolated computation. A student can be fluent in a formula and still miss an AKM question if they can't recognize which formula a real-world scenario calls for.

Teachers already using AI to check numeracy-style questions for accuracy should compare notes across tools first — see Best AI for Math Problems in 2026 (Benchmarked) before trusting any single generated answer key.

Why Indonesia Moved Away From Content-Recall Testing

AN's design isn't arbitrary — it responds directly to a well-documented international benchmark problem, and understanding that rationale helps explain why AKM practice looks so different from old-style review.

Indonesia participates in the OECD's Programme for International Student Assessment (PISA), and the OECD's PISA 2022 results placed the country among the lower-scoring participants in reading and mathematics. That gap was a major part of the policy case for redesigning the national assessment around transferable reasoning rather than syllabus recall, on the reasoning that a test built around applied literacy and numeracy would better reflect — and help build toward — the same skills PISA measures.

• PISA measures similar cognitive layers to AKM: finding information, interpreting it, and reflecting on or evaluating it.
• A curriculum-recall test can't diagnose this gap the way a reasoning-based one can, since a student can pass a content quiz while still struggling to apply the same knowledge to an unfamiliar text or scenario.
• Kurikulum Merdeka's flexibility exists partly to support this shift — giving teachers room to prioritize depth of reasoning over covering every syllabus item at the same pace.

Building AKM-Aligned Habits Into Everyday Lesson Plans

Since only a sample sits AN in any given year, and since the skills it measures are cumulative, the highest-leverage move is building AKM-style reasoning into ordinary lessons across every grade — not just the tested ones — rather than running a standalone test-prep unit.

Grade 5: Reasoning Inside Regular Subjects

At this level, AKM-style habits fit naturally inside subjects a teacher is already covering, without needing a dedicated "literacy period."

1. Pull a short unfamiliar passage into a science or social-studies lesson and ask students to answer one factual and one interpretive question about it.
2. Turn a word problem into a two-step reasoning task — first identify what's actually being asked, then solve.
3. Rotate text types weekly — a diagram one week, a short narrative the next — so students build comfort with varied formats, not just one.

Grade 8 and 11: Cross-Subject Literacy Under Kurikulum Merdeka

Kurikulum Merdeka's flexibility gives secondary teachers real room to fold AKM-style reasoning into any subject, including the Projek Penguatan Profil Pelajar Pancasila (P5) work most schools already run — which connects naturally to the Survei Karakter side of AN.

• Ask students to justify an answer in writing, not just state it — this mirrors AKM's interpret-and-evaluate layer directly.
• Use data from a science or social-studies unit as the basis for a numeracy-style question, rather than treating numeracy as a math-class-only skill.
• Say you teach Grade 8 science: a lesson on population data could double as numeracy practice — reading a table, then reasoning about what it implies — without taking time away from the science content itself.

A Sample Weekly Rhythm Across a Term

A sustainable routine spreads AKM-style practice across the week instead of concentrating it into one long Friday session students dread.

Spreading practice this way also makes each session short enough to fit inside an existing lesson rather than requiring extra class time most schools don't have to spare.

Where AI Genuinely Helps — and Where a Teacher's Judgment Still Leads

Used well, AI shortens the time it takes to build AKM-style practice material without pretending to replace a teacher's read on a specific classroom.

• Generating AKM-style question sets across subjects and text types, once a teacher specifies the grade band and the cognitive layer (find, interpret, or evaluate) they want practiced.
• Drafting varied text types quickly — a short informational passage one day, a simple data table the next — instead of a teacher hunting for suitable unfamiliar material every week.
• Building numeracy word problems tied to real contexts, which is more time-consuming to write from scratch than most teachers have room for weekly.

AI can't see a specific school's actual Rapor Pendidikan results, infer what the Survei Lingkungan Belajar already revealed about that school's climate, or know which text types a particular class has already grown comfortable with — a teacher's ongoing read of the room still drives where to focus next.

• Drafting a rubric for evaluating written justifications — the "explain your reasoning" layer AKM rewards — which many teachers find more time-consuming to write consistently than the questions themselves.
• Translating a passage's difficulty level up or down once a teacher flags that a class found a text too easy or too hard, without rewriting the whole set from scratch.

Handling Student Data Responsibly When Using AI Tools

Generating AKM-style practice for a specific class often means describing that class to an AI tool — grade level, weak areas, sometimes a student's name. Indonesia's Personal Data Protection Law (UU No. 27 Tahun 2022, UU PDP) governs how any platform, including an AI content tool, may collect and process that kind of personal data.

A few habits keep everyday AI-assisted planning on the safer side of that law:

1. Describe a class by grade level and general ability range, not by individual student names, when generating practice material.
2. Favor education-focused tools with a stated data policy over general-purpose chatbots for anything involving specific student performance details.
3. Keep generated practice material and any real student data in separate files, so a shared or exported document never accidentally mixes the two.
4. Check what a free tool does with submitted data before making it part of a weekly routine — a free tier's business model is worth understanding upfront.

None of this blocks using AI for AKM-style practice generation. It just means treating a class description the way any teacher already treats a paper gradebook: useful, but not something to hand to just any tool without checking first.

Tools for Building Asesmen-Nasional-Aligned Practice

No single tool replaces the official ANBK (Asesmen Nasional Berbasis Komputer) practice simulations Kemendikdasmen provides, but a few types of tools each add something different to weekly planning.

You could use EduGenius's class-profile setup to generate a differentiated reading passage and question set at a specific grade band, then adjust the phrasing and context to match what a class has actually studied. Its multi-format export (PDF, DOCX) is designed to make a weekly rotation of fresh, varied practice material genuinely sustainable rather than a once-a-term event.

For a broader look at how AI fits differently into a large-class setting elsewhere in the region, AI for Large Class Sizes in Pakistan covers a structurally different pressure — enrollment size rather than assessment format — worth comparing against.

Related Reading

For how a neighboring country's own bilingual classroom context shapes AI-generated content, see Bilingual AI in Kiswahili and English. Teachers weighing how a mother-tongue-medium classroom changes AI tool selection can also see AI for Teaching in Tamil. For a look at AI applied to a single high-stakes entrance exam rather than a system-wide diagnostic, AI for ECAT and Engineering Entry Tests makes a useful contrast.

For the broader regional picture, see AI in Education Around the World: A 2026 Regional Guide.