The KEAI Compliance Playbook for Law Firms

Indonesia's Kode Etik Advokat Indonesia (KEAI) does not ban legal marketing. It bans two specific things: advertising designed solely to attract attention, and publicity about the outcome of a specific matter. Everything else, including the educational content that drives most legal SEO and GEO strategy, sits inside a permitted zone that PERADI's own leadership has confirmed in public statements. The confusion around what a firm can and cannot publish comes almost entirely from treating the whole code as a blanket restriction rather than reading what the two operative articles actually target.

The Two Articles That Govern Everything

KEAI, established 23 May 2002, contains two provisions with direct implications for digital content. Pasal 8(b) states: "Pemasangan iklan semata-mata untuk menarik perhatian orang adalah dilarang termasuk pemasangan papan nama dengan ukuran dan/atau bentuk yang berlebih-lebihan", prohibiting advertising solely to attract attention, including signage of excessive size or form. Pasal 8(f) states that advocates may not use mass media to seek publicity for themselves or their handling of a matter, except where the disclosure serves to uphold legal principles every advocate is bound to defend. A third provision, Pasal 4(c), separately prohibits promising or guaranteeing a case outcome.

Luhut M.P. Pangaribuan, a code co-drafter and PERADI's own Ketua, has confirmed the rationale publicly, via Hukumonline: the target is not publicity itself but the slippery path from advertising toward promising victories, which Pasal 4(c) already expressly forbids. The restriction is principled, aimed at a specific harm, rather than an absolute prohibition on professional visibility.

Legal Has the Lightest Regulatory Load of the Three Restricted Professions

Comparing KEAI to the advertising rules governing other regulated Indonesian professions clarifies how much latitude actually exists. Notaries, who are public officials under the Undang-Undang Jabatan Notaris (UUJN), face the strictest regime of the three: Article 4 of the Notary Code of Ethics prohibits any self-promotion, advertising or online marketing whatsoever, including sponsoring events or maintaining a promotional profile. Notaries may publish only basic factual contact details, a restriction rooted in their status as public officials appointed by the state rather than private professionals competing for clients, which is precisely why their regime is structurally incomparable to an advocate's. Financial institutions regulated by the Otoritas Jasa Keuangan operate under POJK 22/2023 and POJK 6/2026, which permit active advertising but impose administrative sanctions of up to Rp15 billion for misleading claims or unauthorised influencer content.

Advocates sit in a materially lighter position than either. The core prohibition targets attention-seeking advertising and outcome promises specifically, not commercial visibility itself, and enforcement runs entirely through PERADI's complaint-based Dewan Kehormatan (Honour Council) rather than an administrative-fines regime like OJK's. Compliant, informational SEO and GEO content, built around education rather than persuasion, is structurally unlikely to draw disciplinary attention at all.

The Question Nobody Has Formally Answered

KEAI was written in 2002, a decade and a half before structured data, schema markup or AI-facing content existed as a category. Whether machine-readable metadata, the JSON-LD that tells a search engine or AI system what a page contains, counts as "advertising" or "publicity" under Pasal 8(b) or 8(f) has never been formally addressed by PERADI, any Indonesian court, or the Dewan Kehormatan. Three readings are each independently defensible.

The permissive reading treats schema as machine-readable metadata, not human-facing communication. It cannot "menarik perhatian" in the ordinary sense of the phrase, because no human reads it directly, making it analogous to a library catalogue entry rather than an advertisement. The conservative reading treats the functional effect as identical to advertising regardless of technical form: if schema markup increases the likelihood that an AI system recommends the firm, and that recommendation reaches a potential client, the mechanism causing that outcome is, in substance, promotional. The contextual reading, arguably the most workable, ties the KEAI trigger to commercial and competitive intent rather than technical form: schema describing informational content is permissible because it facilitates access to information, consistent with an advocate's educational duty, while schema specifically engineered to trigger a "best lawyer" recommendation raises the same concern more directly.

PERADI's own 2024 Focus Group Discussion on digital advertising ethics acknowledged that KEAI predates the digital era and requires revision to remain relevant. Until that revision occurs, the operationally defensible position is the contextual reading applied conservatively: implement schema describing informational and educational content, credentials and factual service descriptions, and avoid configurations specifically engineered to win comparative "best" or "top" queries.

Building an Operational Workflow Around the Line, Not Against It

The practical challenge most firms face is not knowing where the line sits. It is that every piece of content gets reviewed as if the line were unclear each time, which turns a workable rule into a bottleneck. The fix, documented consistently across the research on legal content operations, is a structure-first compliance model with four components.

Pre-cleared content templates replace free-form drafting. A template for explaining a licensing process, for example, has rigid sections: statutory facts, administrative steps, mandatory disclaimers, and links to official sources. The reviewing partner's task shifts from assessing a full article line by line to confirming the draft follows the pre-approved structure, which cuts review time from a traditional 10 to 21 day cycle toward 2 to 3 days without lowering the compliance bar.

Modular disclosure blocks are stored centrally and injected programmatically across every relevant page, so a single update to disclaimer language, triggered by a change in PERADI guidance, propagates everywhere at once rather than requiring page-by-page revision. A statutory citation bank holds pre-verified quotes from UU No. 18/2003, KUH Perdata, KUH Pidana and practice-area-specific legislation, checked quarterly against amendments rather than re-verified for every article. Direct statutory quotation, because it cites public regulatory text rather than making an active legal claim, requires only factual citation verification, not a full liability review, and clears the compliance gate faster as a structural consequence of the format itself.

Classifying Every Claim Before Drafting Begins

A practical framework, cross-validated across multiple independent research sources, sorts every proposed claim into one of four categories before a single word is drafted. Factual claims cover office location, qualifications, languages and practice coverage, and clear quickly. Educational claims explain laws, procedures and legal concepts, and represent the bulk of citable, compliant content. Comparative claims, "leading," "top," "most experienced," require the strongest evidentiary support and PERADI's own guidance treats them with particular caution. Outcome-related claims, win rates, guaranteed results, predictions, sit outside the permitted zone entirely and should generally be omitted rather than qualified, since qualification rarely neutralises the underlying Pasal 4(c) exposure.

Applied consistently, this framework does the work a nervous, case-by-case legal review otherwise does inconsistently. A regulatory briefing drafted from the start as an educational claim, with statutory citations pulled from the pre-verified bank and disclaimers injected automatically, arrives at partner review already structured to pass, because the ambiguous decisions were resolved at the classification stage rather than left for the reviewer to untangle from scratch.

What the Workflow Actually Looks Like, Stage by Stage

The difference between the traditional review model and a structure-first one is not a difference in standards. It is a difference in what gets re-examined for every single piece of content versus what gets verified once and reused. Mapped stage by stage, the traditional model spends most of its time re-litigating decisions the structure-first model already settled at the template level.

The total shifts from a 21-day cycle to roughly 2 to 3 days, and the compliance standard applied at each stage does not loosen. What changes is where the judgment gets exercised. Under the traditional model, a senior partner re-decides the same category of question, "does this cross into promotional territory," for every article. Under the structure-first model, that question was answered once, when the template was approved, and every subsequent article inherits the answer.

Why This Matters More for Corporate Practices Than It First Appears

Solo and boutique practitioners tend to publish high volumes of short, transactional content, where a slow review cycle mainly delays a single blog post. Corporate firms operate differently: the content that matters most, regulatory briefings responding to a legislative change, client alerts tied to a live market event, loses most of its value if it is not published within days of the underlying news. A 21-day review cycle does not just slow a corporate firm's content programme. In many cases it makes the highest-value content category functionally impossible to produce on time.

This is also where the fastest-clearing content type, direct statutory and regulator quotation, does the most strategic work. A briefing that quotes the amended regulation's exact text, with the firm's added analysis kept separate and clearly marked as commentary, requires only citation accuracy verification rather than a full liability review, because the firm is reporting public regulatory text rather than making an independent legal claim. That structural distinction is what makes same-week regulatory content realistic without cutting corners on compliance.

Frequently Asked Questions

Can a law firm publish client alerts and regulatory briefings under KEAI?

Yes, and this is one of the clearest permitted categories. UU Advokat Pasal 22 treats legal education as part of an advocate's professional duty, and regulatory briefings explaining a legislative change fall squarely within that duty, provided they stay factual and avoid framing the firm as uniquely superior to competitors.

Does citing a specific matter ever violate Pasal 8(f)?

Discussing an ongoing or completed matter's outcome in mass media generally does, unless the disclosure serves to uphold a legal principle every advocate must defend, the narrow exception the article itself carves out. Anonymised, factual descriptions of matter types handled, without outcome claims, are the safer alternative format used across the research.

Is there a difference between KEAI's restriction and ordinary advertising law?

Indonesia's general electronic-advertising rules independently prohibit inaccurate information and deceptive guarantees, applying to every industry. KEAI adds a profession-specific layer on top: the prohibition on attention-seeking publicity itself, regardless of whether any individual claim is literally false.

What happens if PERADI revises KEAI for the digital era?

Any revision would supersede the current interpretive framework, including the three competing readings on schema and structured data discussed above. Firms following the conservative, informational-only approach are well positioned regardless of which direction a revision takes, since that approach already assumes the stricter reading is the safer default.

Should paid search advertising be avoided entirely until PERADI resolves the grey area?

That is the most defensible current posture. PERADI convened a dedicated 2024 FGD on exactly this question and reached no settled conclusion, which signals genuine institutional uncertainty rather than an oversight a firm can safely read past.

How does a structure-first workflow handle a genuinely novel legal question with no existing template?

It does not try to force one. Pre-cleared templates work for recurring content categories, practice-area explainers, regulatory briefings on familiar legislation types. A genuinely novel question still goes through full review, the same as under the traditional model. The efficiency gain comes from removing the repeat questions from the queue, not from skipping review on the ones that actually need it.

Does a firm need a lawyer to build the statutory citation bank, or can an agency do it?

The citation bank itself, verified quotes from named legislation, can be assembled and quarterly-checked by a content team, but every entry requires sign-off from a licensed advocate before it enters the bank. The agency's role is maintaining the structure and cadence; the substantive legal verification stays with the firm at every stage.

The compliance framework above underlies every recommendation on our Legal SEO and Legal GEO pages, and the same educational-content logic explains why compliant content and AI-citable content converge, a theme covered in more depth in Cited or Silent.