If you run marketing for an Irish public body, you have a statutory advertising quota in Irish, the largest bodies are collectively missing it, and that is a compliance problem with a content solution.
If you run marketing for an ordinary private Irish company, you have no such obligation, there is no measured commercial search demand for Irish-language content in your category, and any agency selling you a bilingual site on the basis of national sentiment is selling you something the evidence does not support.
Both of those statements are true at the same time. Most discussion of Irish-language content collapses them into one, which is how public bodies end up under-informed about a real legal duty and private companies end up buying content nobody searches for.
What Section 10A Actually Requires
The obligation comes from Section 10A of the Official Languages (Amendment) Act 2021. It sets two separate numeric thresholds for public bodies, and they are frequently conflated with each other.
These apply to public bodies. Final provisions, sections 4, 5 and 17, commenced on 21 December 2024.
Of advertising must be in Irish
A proportion of the total advertising a public body places annually has to be in the Irish language.
Of advertising budget to Irish-language media
A proportion of the money, not the volume, has to go to media outlets where half or more of the content is in Irish.
Compliance reality
An Coimisinéir Teanga's reporting shows that Category A public bodies, those spending above EUR 750,000 on advertising, on average met neither threshold. Sector spending of EUR 3.8 million on Irish-language media equated to roughly 4% of budget against the 5% requirement. Category B bodies, spending between EUR 250,000 and EUR 749,000, performed better.
Note the asymmetry between the two thresholds. The 20% figure counts advertising, meaning volume of placements or campaigns. The 5% figure counts money. A body could theoretically satisfy one and fail the other by a wide margin, and the compliance data suggests many are failing both by different amounts.
The definition of Irish-language media is also narrower than people assume. It is not an outlet that occasionally publishes in Irish. It is an outlet where half or more of the content is in Irish. That is a much smaller list of eligible destinations, and it constrains media planning in a way that a percentage target alone does not communicate.
Who This Applies To, and Who It Does Not
| Organisation type | Section 10A obligation | What this means for content planning |
|---|---|---|
| Public bodies Government departments, agencies, local authorities, State bodies | Yes. Both the 20% advertising and 5% budget thresholds apply. | Irish-language content is a compliance deliverable with a measurable target, not a discretionary brand choice. It needs a named owner and an annual reporting position. |
| Contractors delivering public-facing services Third parties acting on a public body's behalf | Potentially. The obligation can pass through where the contractor delivers services to the public for the body. | Worth checking your contract language before assuming it does not reach you. This is the part most often missed on both sides of a public procurement. |
| Private commercial businesses Ordinary Irish companies with no public service function | No. Section 10A does not apply. | Any Irish-language content is a commercial decision that has to justify itself on commercial grounds. There is no legal driver. |
| Foreign companies selling into Ireland | No, unless contracting to deliver public services. | Same as above. English-language content correctly localised for Irish jurisdiction, currency and regulators does more for you than a bilingual site. |
The contractor row is the one worth pausing on. A private company that wins a contract to deliver a public-facing service on behalf of a public body can inherit language obligations it never considered during the tender. If you are bidding for Irish public sector work, that clause belongs on your compliance checklist rather than being discovered after award.
Two Twenty Percents That Are Not the Same Thing
A small but consequential confusion. There is a separate 20% figure in Irish language policy: the target that 20% of new recruits to the public service should be competent in Irish by 2030.
That is a workforce target, not an advertising target. It has nothing to do with Section 10A media obligations. I have seen the two merged in briefing material more than once, usually producing a confident but incorrect statement that public bodies must publish 20% of content in Irish by 2030. They must place 20% of advertising in Irish now, and separately aim for 20% Irish-competent recruitment by 2030. Different obligations, different mechanisms, different deadlines.
The Commercial Case, and Why It Rests on One Survey
Anyone arguing for Irish-language content on commercial grounds will eventually reach for the same research. It is worth examining rather than repeating.
A survey of 1,000 respondents conducted in February 2024 by Amárach Research for Gaelchultúr, reported by the Irish Times on 4 March 2024, found that 41% said they would be more likely to buy from a company using the Irish language, 73% assumed a product using Irish was locally made, and 26% said they would be willing to pay more.
Those are respectable numbers. Three caveats belong with them every time they are cited.
It is a single source. No corroborating study was located across four independent research passes. A finding with one source is a finding, not a consensus.
It is stated preference, not observed behaviour. Survey respondents saying they would be more likely to buy is a different measurement from anyone actually buying. The gap between stated and revealed preference on questions touching national identity is well known and tends to run in one direction.
And it is from early 2024. In a period when everything about search behaviour has been changing, a survey more than two years old should be dated explicitly rather than presented as current sentiment.
What would strengthen the commercial case is search demand data, and this is where the argument runs out. Across four independent research passes, no measurement of private-sector commercial demand for Irish-language content was located, whether expressed as search volume, enquiries or revenue. Not a small number. No number.
That absence is not proof the demand is zero. It does mean nobody has demonstrated it exists at commercially meaningful scale, and the burden of proof sits with whoever is proposing the spend.
What Happens When Someone Prompts an Assistant in Irish
This is the newest version of the question and the one with the least evidence behind it. It is also the one where the two plausible answers point in opposite directions, which is worth setting out honestly rather than picking whichever suits a proposal.
No study measures citation behaviour for prompts written in Irish. All four independent research passes conducted for this article returned the same verdict.
Prompting in Irish strengthens Irish source retrieval
Research into citation behaviour across languages, published August 2026, put 400 questions to four AI systems across five countries in both English and the local language. French-language queries sent 16% of citations to French domains against 1.4% for English queries. Japanese-language queries sent 26% to Japanese domains against roughly 1% in English.
The sub-finding matters more for public bodies than the headline. French government domains went from near-zero in English to 1.25% in French, and Japanese government and academic domains from 0.13% in English to 4.26% in Japanese. Official-sector content specifically gained the most from the local-language query.
If that pattern generalises, an Irish-language prompt would pull Irish-hosted, Irish-authored sources into the answer far more reliably than the English equivalent, because the Irish-language corpus is almost entirely Irish by definition.
The corpus is too small to produce a good answer
The same mechanism cuts the other way at low volume. The Irish-language web is orders of magnitude smaller than the English-language web, so an Irish-language prompt may retrieve from a very thin pool.
The result could be lower source diversity, weaker factual grounding, more reliance on translation, or an assistant simply answering in Irish from English sources without citing anything Irish at all.
For a public body with statutory Irish-language obligations, testing this is worth doing, because you are producing the content anyway and the question of whether assistants surface it is a legitimate performance question. A modest prompt panel in Irish, run alongside the English one, answers it for your own domain in a couple of weeks.
For a private company, it is not a priority. You would be commissioning content with no measured demand in order to test a hypothesis nobody has resolved, in a language your buyers are probably not searching in.
What We Would Actually Tell a Private Company
Direct answer: in almost all cases, do not build Irish-language content, and put the same budget into jurisdictional accuracy instead.
Here is why that trade is better. Irish English follows British spelling conventions, so spelling signals nothing about which market a page serves. What does signal it is everything underneath: euro pricing with VAT treatment stated, the Central Bank of Ireland rather than the FCA, the Data Protection Commission rather than the ICO, the Companies Registration Office rather than Companies House, Revenue rather than HMRC, Eircodes rather than postcodes, solicitors rather than attorneys, county and city coverage that matches real service areas.
Every one of those is a signal a retrieval system can verify and a buyer can recognise. Together they do more to establish that a page is genuinely about the Republic of Ireland than a translated homepage ever will. That is the substance of how we approach SEO for the Irish market, and it is also why we tell clients when Irish-language content is not the answer.
Three exceptions where the calculus changes. A business whose customer base genuinely sits in a Gaeltacht area. A brand whose positioning is explicitly built on Irish cultural identity, where the language is the product rather than a channel. And any organisation bidding for public sector contracts, where the contractor pass-through may create an obligation regardless of commercial logic.
What Public Bodies Should Do
Five things, in order.
Establish which category you fall into. Above EUR 750,000 in annual advertising spend puts you in the group that the compliance reporting shows performing worst against both thresholds. That is a reason for scrutiny, not comfort in numbers.
Track the two thresholds separately. Volume of advertising in Irish, and percentage of budget to qualifying Irish-language media. They are different measurements and a single dashboard number will hide a failure on one of them.
Verify that your chosen outlets qualify. The statutory test is 50% or more of content through Irish. An outlet with an Irish-language section may not meet it, and spend placed with a non-qualifying outlet does not count toward the 5%.
Build the Irish-language requirement into the content calendar, not the sign-off gate. Retrofitting Irish content at the end of a campaign cycle produces translation rather than communication, and it produces it late.
Check your contractors. If third parties deliver public-facing services on your behalf, the obligation may extend to their output, and their contracts should say so explicitly rather than leaving it to interpretation.
What Nobody Has Measured
Four gaps, stated plainly, because this is a topic where confident assertion is common and evidence is thin.
There is no measurement of private-sector commercial demand for Irish-language content, in search volume, enquiries or revenue. Three independent research passes returned unavailable.
There is no measurement of citation behaviour for prompts written in Irish. Four passes returned unavailable.
There is no corroboration of the Amárach consumer sentiment figures from a second independent study, and the original is from early 2024.
And there is no published Irish study of whether Irish-language content improves organic performance for public bodies, which is a strange gap given the obligation now has statutory force and a measurable target attached.
The honest position on all of this is narrower than either side of the usual argument. The legal duty is real, specific and currently unmet at the top of the public sector. The commercial case for private companies rests on one dated survey of stated preference and no demand data at all. And the AI question is genuinely open, with two credible hypotheses pointing opposite ways and no test performed. Anyone telling you otherwise, in either direction, is filling a gap with confidence. The same discipline applies to how we handle AI visibility measurement for Irish clients, where the temptation to assert unmeasured things is even stronger.
The broader framework for deciding what to publish when the demand signal is missing is set out in Found Before They Search, and the gated edition is free to download. Our cross-market measurement work is published in the AI Citation Rate Report 2026.
Written by Tessar Napitupulu, Founder and CEO of PT Arfadia Digital Indonesia, a member of the Forbes Agency Council, and author of Found Before They Search and Cited or Silent. This article describes statutory obligations for general information and is not legal advice. Irish public bodies and their contractors should confirm their own position with an Irish-qualified practitioner.
Frequently Asked Questions
Does my Irish website need Irish-language content?
Almost certainly not, unless you are a public body or you deliver public-facing services on behalf of one. Section 10A of the Official Languages (Amendment) Act 2021 imposes advertising obligations on public bodies, not on private commercial businesses. For an ordinary Irish company there is no legal driver, and across four independent research passes no measurement of private-sector commercial demand for Irish-language content could be located in search volume, enquiries or revenue. The better use of the same budget is jurisdictional accuracy in English.
What exactly does Section 10A require?
Two separate thresholds for public bodies. At least 20% of the advertising a public body places annually must be in the Irish language, and at least 5% of its advertising budget must go to Irish-language media. The two count different things, one counting advertising volume and the other counting money, so a body can satisfy one while failing the other. Final provisions, sections 4, 5 and 17, commenced on 21 December 2024.
What counts as Irish-language media?
The statutory test is media where 50% or more of the content is through Irish. That is narrower than it sounds and is frequently misread. An outlet that publishes an Irish-language section alongside mostly English content does not qualify, and budget placed with a non-qualifying outlet does not count toward the 5% requirement. No further statutory guidance exists beyond that definition, which makes verifying eligibility a practical step rather than an assumption.
Are public bodies actually meeting these thresholds?
The largest ones collectively are not. An Coimisinéir Teanga's reporting shows that Category A public bodies, those spending above EUR 750,000 on advertising annually, on average met neither the 20% nor the 5% threshold. Sector spending of EUR 3.8 million on Irish-language media equated to roughly 4% of budget against the 5% requirement. Category B bodies, spending between EUR 250,000 and EUR 749,000, performed better. Being in the larger group is a reason for scrutiny rather than reassurance.
Can the obligation reach a private contractor?
Yes, potentially. Where a third party delivers public-facing services on behalf of a public body, the language obligation can pass through to that contractor's output. This is the part most often missed on both sides of a public procurement, and it should appear on a compliance checklist during the tender rather than being discovered after award. Contract language on this point is worth reading explicitly rather than interpreting.
Is the 20% advertising target the same as the 20% recruitment target?
No, and merging them is a common error. Section 10A requires that at least 20% of a public body's advertising be in Irish, and that obligation applies now. Separately, Irish language policy sets a target that 20% of new recruits to the public service should be competent in Irish by 2030. One is a media obligation, the other a workforce target. Different mechanisms, different deadlines, and no relationship between them beyond the coincidence of the number.
Is there commercial evidence that using Irish helps sales?
There is one survey, and it should be cited with its limitations attached. Amárach Research surveyed 1,000 respondents for Gaelchultúr in February 2024, reported by the Irish Times on 4 March 2024, finding that 41% said they would be more likely to buy from a company using Irish, 73% assumed a product using Irish was locally made, and 26% would be willing to pay more. Three caveats: it is a single source with no corroborating study located, it measures stated preference rather than observed purchasing behaviour, and it is from early 2024. No search demand data supports it.
Do AI assistants behave differently for prompts written in Irish?
Nobody has measured it. All four independent research passes conducted for this article returned unavailable on this question. Two credible hypotheses point in opposite directions. Research published in August 2026, putting 400 questions to four AI systems across five countries in both English and the local language, found 16% of citations going to French domains on French-language queries against 1.4% in English, and 26% to Japanese domains on Japanese queries against roughly 1% in English. Official-sector domains gained most: French government sites rose from near-zero in English to 1.25% in French. That suggests Irish-language prompting could strengthen Irish source retrieval, which is directly relevant to public bodies. Against that, the Irish-language web is orders of magnitude smaller, so the same mechanism could produce thin retrieval, weaker grounding, or an answer in Irish drawn entirely from English sources. Both are reasonable. Neither has been tested.
Should a public body test Irish-language AI visibility?
It is worth doing, because you are producing the content anyway under a statutory obligation and whether assistants surface it is a legitimate performance question. A modest prompt panel in Irish run alongside the English one answers it for your own domain within a couple of weeks. For a private company the same exercise means commissioning content with no measured demand in order to test a hypothesis nobody has resolved, which is a much weaker case.
If not Irish-language content, what should we invest in instead?
Jurisdictional accuracy in English. Irish English follows British spelling conventions, so spelling signals nothing about which market a page serves. What does signal it is the substance underneath: euro pricing with VAT treatment stated, the Central Bank of Ireland rather than the FCA, the Data Protection Commission rather than the ICO, the Companies Registration Office rather than Companies House, Revenue rather than HMRC, Eircodes rather than postcodes, solicitors rather than attorneys, and county coverage matching real service areas. These are verifiable by a retrieval system and recognisable by a buyer, which a translated homepage is not.
Are there exceptions where Irish-language content does make commercial sense?
Three. A business whose customer base genuinely sits within a Gaeltacht area, where the language is the working language of the market rather than a gesture. A brand whose positioning is explicitly built on Irish cultural identity, where the language is part of the product rather than a distribution channel. And any organisation bidding for public sector contracts, where contractor pass-through may create an obligation independent of commercial logic.
Sources & References:
- Official Languages (Amendment) Act 2021, Section 10A. Requires at least 20% of advertising placed annually by a public body to be in the Irish language, and at least 5% of its advertising budget to be spent on Irish-language media, statutorily defined as media where 50% or more of content is through Irish. No further statutory guidance exists beyond that definition.
- Government of Ireland, announcement of 17 December 2024 on the commencement of the final provisions of the Act, sections 4, 5 and 17, with effect from 21 December 2024.
- An Coimisinéir Teanga, reporting on Section 10A compliance. Category A public bodies, with advertising spend above EUR 750,000, on average met neither the 20% nor the 5% threshold. Sector spending of EUR 3.8 million on Irish-language media equated to approximately 4% of budget against the 5% statutory requirement. Category B bodies, spending between EUR 250,000 and EUR 749,000, performed better.
- Institute of Advertising Practitioners in Ireland, guidance material on Section 10A obligations for advertisers and agencies working with public bodies.
- Separate Irish language policy target that 20% of new recruits to the public service be competent in Irish by 2030. A workforce target, distinct from and unrelated to the Section 10A advertising obligations, included here because the two are frequently conflated.
- Amárach Research for Gaelchultúr, survey of 1,000 respondents conducted February 2024, reported by the Irish Times on 4 March 2024. 41% said they would be more likely to buy from a company using the Irish language, 73% assumed a product using Irish was locally made, 26% said they would be willing to pay more. REPORTED, single source, stated preference rather than observed behaviour, and dated early 2024. No corroborating study located.
- Weglot, "How AI Search Engines Cite Sources by Language and Country", published August 2026. Methodology: 400 questions across 5 topic categories and 3 levels of specificity, written in English, French, Spanish and Japanese, put to 4 AI systems across 5 countries, producing more than 16,000 answers with every cited source logged. French-language queries sent 16% of citations to .fr domains against 1.4% for English queries, with French government .gouv.fr domains rising from near-zero in English to 1.25% in French. Japanese-language queries sent 26% of citations to .jp domains against approximately 1% in English, with Japanese government and academic .go.jp and .ac.jp domains at 4.26% in Japanese against 0.13% in English. Spanish-language queries in Spain sent 7% of citations to .es domains against approximately 1% in English. Ireland and the Irish language were not among the cases tested.
- Ahrefs Brand Radar, "Which Countries Have the Most AI Overviews", published 4 November 2025, 108 million AI Overview queries. English accounted for 52.75% of all AI Overviews worldwide. Irish does not appear in the measured language table.
- Private-sector commercial demand for Irish-language content, measured as search volume, enquiries or revenue: UNAVAILABLE. Three independent research passes returned no measurement.
- AI citation behaviour for prompts written in Irish: UNAVAILABLE. Four independent research passes returned no measurement.
- Effect of Irish-language content on organic search performance for public bodies: UNAVAILABLE. No published Irish study located.