BahanCek — self-declare prep engine for Indonesian halal companions

1. One-liner

Reads a UMKM ingredient label, flags the syubhat additives, and drafts the SIHALAL self-declare dossier a halal companion signs.

2. Trend signal — why now?

There is a hard, government-set fuse burning: 17 October 2026. On that date Indonesia’s BPJPH begins mandatory enforcement of halal certification for food, beverage, and slaughter products made by micro and small enterprises (UMKM). Non-certified products get administrative warnings, pulled from retail shelves, delisted from e-commerce, and fined. Enforcement (active market supervision) begins the very next day, 18 October 2026.

The gap is enormous. From 2019 to mid-2024, BPJPH issued ~4.4M halal certificates — only ~44% of a 10M target. Even the more optimistic count puts the country at ~6.4–7M certified products by end of 2025 against millions of uncertified UMKM SKUs that must clear the line in the next fifteen months. BPJPH has itself publicly warned that a surge in applications near the deadline will increase processing times and the risk of bottlenecks.

The free path — the SEHATI program’s self-declare scheme (1.35M free certificate quotas for 2025–2026) — is the route almost every micro food seller will take. But self-declare is not automatic. It requires a PPH companion (Pendamping Proses Produk Halal) to verify and validate the applicant’s ingredient declaration and written production procedures before it goes to the MUI fatwa sitting. There are only ~54,164 registered companions under ~187 LP3H institutions, most of whom do this part-time around a day job. The reported failure modes are consistent: UMKM assume “no pork = halal” and miss critical-point ingredients (emulsifiers, flavorings, animal-derived additives); they keep no traceable raw-material documentation; they have no written procedures; they fat-finger the SIHALAL data entry. Every one of those causes a rejection or a delay — and the companion eats the rework.

Provenance:

• Signal 1 (Demand): Millions of UMKM food SKUs still uncertified vs. a 10M target; hard enforcement deadline 17 Oct 2026 with delisting + fines; self-declare rejections driven by ingredient-tracing and missing procedures — https://en.antaranews.com/amp/news/408710/indonesia-to-enforce-mandatory-halal-certification-in-october-2026 and https://www.peluanghalal.id/blog/kesalahan-umum-pengajuan-sertifikat-halal-self-declare-umkm-dan-cara-menghindarinya — 2026-07-03
• Signal 2 (Feasibility): The critical reference data is government-published and structured — KMA 1360/2021 “positive list” of 4,038+ exempted materials — so a Bahasa-Indonesia LLM + label OCR can classify ingredients against a known rulebook rather than a proprietary dataset — https://cmsbl.halal.go.id/uploads/KMA_1360_Tahun_2021_tentang_Bahan_yang_Dikecualikan_dari_Kewajiban_Bersertifikat_Halal_a6401618a5.pdf — 2026-07-03
• Signal 3 (Economic): ~54,164 companions for millions of applicants; BPJPH disbursed Rp81B in incentives to companions/LP3H; private “jasa pengurusan” consultants already charge Rp15M+ for the same dossier prep — money is clearly moving in this workflow — https://kemenag.go.id/nasional/bpjph-cairkan-rp81-m-insentif-pendamping-proses-produk-halal-dan-lp3h-sZRPx and https://jasanotaris.co.id/panduan-lengkap-biaya-sertifikasi-halal-untuk-umkm-dan-perusahaan/ — 2026-07-03 Category: Regulatory arbitrage

3. The opportunity

Self-declare is free but labor-bound. The cost isn’t the Rp0 certificate — it’s the hours of human prep: reading each product’s ingredient list, tracing every additive to a halal-guaranteed source, deciding whether an item is exempt (positive list), safe, or syubhat (doubtful, needs documentation), writing the production procedure, and hand-keying it all into SIHALAL without error. Today that work is done by a scarce, part-time companion, one product at a time, or by a Rp15M consultant a micro-seller can’t afford.

The incumbent to disrupt is the manual companion workflow itself and the overpriced consultant tier above it. A companion who can currently prep and validate ~2–4 products in an evening should be able to prep 15–20 if the ingredient classification and the dossier drafting are done for them and they just review-and-sign. That’s a 5× throughput jump on the exact bottleneck BPJPH says will jam near the deadline. The rulebook is public (KMA 1360/2021 positive list + known critical-point categories), so this is a classification-against-a-known-standard problem, not a “we need proprietary halal data” problem — which is precisely why it’s buildable by a small team.

4. Target market

• Primary customer: Active PPH companions and the LP3H institutions that manage them — the ~54,164 registered companions and ~187 LP3H, concentrated in Java (Jakarta, Surabaya, Bandung, Semarang) and the larger provincial capitals. Secondary: private halal-certification consultancies (“jasa pengurusan sertifikat halal”) who prep dossiers for a fee and want to do more volume per staffer.
• Why they buy, in their words: “Banyak UMKM tidak punya dokumentasi bahan yang rapi” (most UMKM have no orderly ingredient documentation) and “keterbatasan waktu pendamping yang sebagian besar punya pekerjaan utama” (companions have limited time because most have a main job). They lose evenings to rework caused by applicants who declared an ingredient wrong. They’re paid per certificate — so faster prep is directly more income.
• Rough TAM reasoning: Even a serviceable slice — say 5,000–10,000 active, high-volume companions and 300–500 consultancies willing to pay for a tool that multiplies their per-certificate income during the 2025–26 crunch — is a large enough beachhead for a sub-$5M ARR business. Millions of UMKM SKUs still need dossiers before 17 Oct 2026.
• Why now for them: The deadline is fixed and enforcement is imminent. Volume is about to spike beyond what part-time humans can clear by hand, and BPJPH has publicly said so. The pain moves from “annoying” to “I physically cannot process my queue” in Q3–Q4 2026.

5. Product sketch (MVP)

• Snap or upload a product’s ingredient label / supplier spec / MSDS — BahanCek OCRs it in Bahasa Indonesia and extracts the ingredient list.
• Each ingredient is auto-classified into one of three buckets: exempt (matches the KMA 1360/2021 positive list), halal-guaranteed (matches an existing halal-certified material), or syubhat / critical point (emulsifier, flavoring, gelatin, enzyme, animal-derived, alcohol-adjacent — needs a source document).
• For every syubhat item, a plain-language checklist of exactly what the companion must ask the UMKM for (e.g. “get the supplier’s halal certificate for this emulsifier” / “confirm the mono-and-diglyceride source is plant”).
• Auto-drafts the written PPH procedure documents (raw-material list, production process, equipment-separation statement) in the format the self-declare submission expects.
• A pre-submission validation pass that flags the common rejection triggers — mismatched product name, missing composition field, an ingredient with no traceable source — before it ever reaches SIHALAL.
• A companion dashboard: queue of UMKM in progress, which are dossier-ready, which are blocked on a missing supplier document.
• Bahasa-Indonesia throughout; simple enough for a low-digital-literacy applicant to feed photos in over WhatsApp and for the companion to review on a phone.

6. AI angle — what’s load-bearing

Remove the AI and this is a PDF checklist nobody uses. The load-bearing work is: (1) OCR + extraction of messy, photographed, multi-language ingredient labels into a clean ingredient list; (2) classification of each ingredient against the positive list and the critical-point taxonomy — including fuzzy matching of trade names and synonyms to their underlying material, which is where humans miss the syubbat items; and (3) generation of the written procedure documents from a few structured inputs. The rulebook is deterministic and public, but the reading and matching of real-world labels to that rulebook is exactly the judgment-heavy step that eats a companion’s evening and causes rejections. That’s the AI’s job. The human companion stays in the loop as the required legal validator — the AI never claims the halal status, it preps the dossier the companion signs.

7. Localization angle

This is the localization play — it does not exist outside Indonesia’s specific regime. It’s Bahasa-Indonesia-first, built around SIHALAL’s exact submission fields, the KMA 1360/2021 positive list, the self-declare eligibility criteria (Rp500M turnover cap, non-high-risk products), and the PPH companion role. Distribution is WhatsApp-native because that’s how UMKM and companions already communicate. Pricing is in rupiah at a tier a per-certificate-paid companion can absorb (a fraction of the Rp15M a private consultant charges). A generic global “compliance” tool cannot touch this — the value is entirely in the local rulebook and workflow.

8. Business model — path to $1M–$5M ARR

• Pricing: Two tiers. (a) Companion / consultant subscription — Rp299,000–499,000/month (~$18–30) per seat for unlimited dossier prep, aimed at high-volume companions and LP3H staff. (b) Per-dossier credits — Rp25,000–50,000 per completed product dossier for occasional users, sold in packs. LP3H institutions buy seat bundles for their companion pools.
• ACV: A committed companion/consultant subscriber ≈ Rp4.8M/year (~$300). LP3H bulk seats and heavy consultancies push higher via multiple seats.
• Rough math to $1M ARR: ~$1M ≈ Rp16B. At ~$300 ACV that’s ~3,300 paying seats — a small fraction of 54,000+ companions plus consultancies, entirely reachable during the 2025–26 crunch. Or a blended mix of ~2,000 subscriptions + several hundred thousand per-dossier credits.
• Rough math to $5M ARR: Needs ~15,000+ seats or a dominant position selling per-dossier credits to LP3H at institutional scale, plus expansion into cosmetics/medicine self-declare (same mechanic, different rulebook) which come under the same 2026 mandate.
• Expansion path: Add cosmetics, drugs, and consumer-goods rulebooks (all mandated 2026); sell a white-label LP3H edition; add a renewal/monitoring tier (certificates expire and product formulations change, so re-prep is recurring beyond the one-time deadline rush).

9. Go-to-market wedge — first 100 customers

• LP3H institutions are a list, not a mystery. BPJPH publishes registered LP3H (there are ~187). Approach the 30–40 largest by companion count directly with a “clear your Q3 queue 5× faster” pitch and a free trial on their next 20 dossiers. Each LP3H that says yes brings dozens of companion seats.
• Companion WhatsApp groups. PPH companions coordinate in provincial WhatsApp/Telegram groups. Recruit 10–20 high-volume companions as design partners, give them free access through the deadline, and let per-certificate-income math do the word-of-mouth (more dossiers/evening = more pay).
• Deadline-panic content. Short Bahasa reels/TikToks aimed at UMKM: “kenapa self-declare halal kamu ditolak” (why your self-declare got rejected) driving UMKM to pressure their companion — and to the per-dossier tier directly.
• Piggyback SEHATI events. BPJPH and universities run mass certification drives (“300 titik” companion capacity events). Sponsor/attend as the prep tool the companions use on the day.
• Consultant channel. The Rp15M “jasa pengurusan” consultancies want margin — sell them a tool that lets one staffer do 5× the dossiers.

10. Build complexity — justification

Medium. The reference data is off-the-shelf and public (positive list, self-declare criteria, SIHALAL field spec), and the core is off-the-shelf OCR + a Bahasa-Indonesia LLM classifying against a known rulebook and generating templated documents. The genuinely non-trivial work is (a) building and maintaining an accurate ingredient-synonym/critical-point knowledge base so classification is trustworthy, and (b) the WhatsApp-first ingestion for low-digital-literacy users. No SIHALAL API integration is required for v1 — the companion still does the final human submission, which is legally necessary anyway. A small team ships a credible v1 in ~8–12 weeks.

11. Gating checklist

Gate	Pass?	Note
Legal in target market	✅	Tool assists dossier prep; a licensed PPH companion remains the legal validator. No claim of halal status by the software.
Ethical — no harm / dark patterns	✅	Reduces rejection risk and cost for micro-sellers; keeps the human companion in the loop. Must be careful not to over-claim certainty on syubhat items.
Market exists (evidence above)	✅	Millions of uncertified SKUs, hard deadline, scarce companions, existing paid consultants.
1–5 person team can build this	✅	Off-the-shelf AI + public rulebook + templated docs.
Launchable with <$50K / ₹40L	✅	No hardware, no license, no proprietary data buy.

12. Feasibility score

Axis	Weight	Score	Notes
Problem intensity	20	16/20	Real, deadline-forced, repeated pain — but it’s the companion’s pain to buy away, and part-timers may just grind through it manually.
Demand evidence	15	13/15	Multiple hard signals: enforcement date, backlog vs target, documented rejection causes, existing paid consultants, BPJPH’s own bottleneck warning.
Build feasibility	15	11/15	Public rulebook helps a lot; the ingredient-synonym knowledge base and OCR-on-messy-labels are the real work. ~8–12 weeks.
Distribution clarity	15	12/15	LP3H list + companion WhatsApp groups are concrete named channels; conversion during the crunch is favorable.
Revenue mechanics	15	11/15	Rupiah pricing works; but per-certificate-paid companions are price-sensitive and the acute demand is time-boxed to the deadline.
Time to first revenue	10	8/10	Deadline urgency + a tool that prints income for companions = fast paid pilots, weeks not months.
Defensibility	10	5/10	Public rulebook means copyable; moat is the maintained ingredient knowledge base, LP3H relationships, and speed. Post-deadline demand softens.
Total	100	76/100

13. Qualitative modifiers

Founder-fit tags

operations-heavy · domain-expertise-required — you need someone who genuinely understands the SIHALAL self-declare workflow and can build LP3H relationships on the ground in Indonesia; the tech is the easy half.

Key assumptions to validate (3–5)

1. Assumption: Companions/LP3H will pay a monthly rupiah subscription (not expect it free like SEHATI). How to test: Put a price in front of 30 high-volume companions and 5 LP3H directors; measure willingness to pre-commit for a trial.
2. Assumption: The AI can classify real, photographed UMKM ingredient labels against the positive list + critical points accurately enough that companions trust it (low false-”safe” rate). How to test: Blind-test on 200 real product labels vs. an expert companion’s judgment; target >90% agreement, near-zero missed syubhat items.
3. Assumption: The acute demand isn’t a one-time spike that evaporates on 18 Oct 2026. How to test: Model recurring demand — new UMKM formation rate, certificate expiry/renewal, formulation changes, and the cosmetics/drug categories under the same mandate.
4. Assumption: 5× throughput is real, not theoretical. How to test: Time a companion doing 5 dossiers by hand vs. with the tool; measure minutes saved and rejection-rate delta.

Risk flags

1. Market timing / demand cliff: The sharpest demand is tied to the 17 Oct 2026 deadline. If enforcement is delayed again (it has been extended before) or demand collapses after the rush, the recurring base must come from renewals + adjacent categories. Build for that from day one.
2. Regulatory dependency: BPJPH/SIHALAL owns the process; a change to the self-declare rules, forms, or eligibility could force rapid rework. Also risk that BPJPH ships its own free tool.
3. Accuracy / trust liability: A wrongly-classified “safe” ingredient that later fails audit damages the companion and the brand. Positioning must be assist-not-decide, human-in-the-loop, with conservative flagging.
4. Price sensitivity: Companions are paid modestly per certificate and UMKM are micro. Pricing has to sit well below the Rp15M consultant tier and be justified by income earned, not features.

14. Structured verdict

Score:                  76/100
Verdict:                GO
Confidence:             Medium
Best-fit builder:       Indonesia-based operator with halal/SIHALAL domain knowledge + a technical co-founder; strong LP3H/companion relationships
Time to revenue:        4–8 weeks (deadline urgency drives fast paid pilots)
Capital to launch:      ₹8–15 lakh / ~$10–18K
Top 3 assumptions to validate first:
  1. Companions/LP3H pay a rupiah subscription — price 30 companions + 5 LP3H directors, measure pre-commit
  2. Classification accuracy on real photographed labels >90% with near-zero missed syubhat — blind test on 200 labels
  3. Demand survives past 17 Oct 2026 via renewals + cosmetics/drugs categories — model recurring base
Kill criteria:
  - Abandon if <20% of 40 high-volume companions will pay after a free trial through one real deadline cycle
  - Abandon if AI misses syubhat ("safe" false-negative) on >2% of a 200-label blind test — trust is the product
  - Abandon if BPJPH ships an equivalent free prep tool inside SIHALAL before your v1 lands

15. Next step — 1-week validation sprint

• Day 1–2: Pull the public LP3H list and join 3–4 provincial PPH companion WhatsApp groups. Line up 15–20 interviews. Collect 200 real photographed ingredient labels from actual UMKM products.
• Day 3–4: Build a throwaway prototype — OCR + LLM classifying those 200 labels against the KMA 1360/2021 positive list and a hand-built critical-point list. Have one expert companion grade the output. Simultaneously, put a rupiah price in front of interviewees and ask for a trial pre-commit.
• Day 5: Decide. Go only if: (a) ≥90% classification agreement with near-zero missed syubhat items on the 200-label test, AND (b) ≥8 of 20 companions verbally commit to a paid trial at the proposed price. Both are falsifiable numbers — no vibes.