Israel's $1.58 Billion Ad Market Is Being Rewritten Inside ChatGPT

Thirty-five percent of consumers now start product research inside an AI tool. Search engines get 13.6%. In B2B, 42% of decision-makers open the buying process with a query to a large language model. This is not a forecast. It is a measured change in behavior, and it has already arrived in the Israeli market — a $1.58 billion digital ad economy running on two platforms that are being converted into answer engines in real time.

Israeli banks, telecoms, HMOs, tech exporters, and airlines are competing for the same thing they always were — the customer's first question. The customer just stopped asking Google.

The economics moved. The playbook has to move with them.

Muck Rack analyzed more than one million AI prompts. 85.5% of the citations inside AI answers come from earned media, not brand-owned content. University of Toronto research puts the ratio at roughly five to one. Broad distribution — the boring, unsexy work of syndication and third-party coverage — lifts AI citation rates by up to 325%. A brand present on four or more third-party platforms is 2.8 times more likely to be cited than a brand relying on its own domain.

Translation for CFOs: the retrieval layer of the AI economy is fed by earned media. The paid-media playbook that dominated the last decade of Israeli marketing is a secondary input to the answer the buyer actually sees.

AI referral traffic converts at 14.2%. Google organic converts at 2.8%. Five to one. The volume is smaller today. The quality is not comparable.

Israel is unusually exposed — in a good way

Five structural features make the Israeli market one of the most concentrated GEO opportunities in the developed world.

Platform concentration. Google captures roughly 46% of Israeli digital ad spend. Meta captures another 15%. Sixty-one percent of the local digital economy flows through two platforms that are integrating AI answers directly into the surface the buyer already looks at.

Oligopoly structure. Five major banks. Three cellular carriers. Four HMOs. Two dominant supermarket chains. In an oligopoly, share of model maps directly onto share of wallet. A fragmented market absorbs AI-visibility gaps. A concentrated one does not.

The mortgage and credit cycle. Israeli banking sector assets grew 10 to 12% year-over-year through 2024 and 2025 on the back of mortgage origination. High-intent Hebrew financial queries are among the most valuable uncontested pieces of AI real estate in any market anywhere. Nobody owns them yet.

Export dependency. The largest Israeli corporates derive significant revenue from overseas. Their visibility inside English-language AI answers is now a direct input to export demand. English and Hebrew are two different markets that have to be worked independently.

High digital engagement. Israeli consumers index above OECD averages on smartphone penetration, e-commerce adoption, and digital service use. AI behaviors arrive here earlier and faster than in peer markets. The transition curve is compressed.

The Hebrew arbitrage window

Hebrew is a minority training language in every frontier AI model. The Nagel Committee's August 2025 report to the Israeli government flagged the absence of a Hebrew national language model as a digital sovereignty issue. It is also a commercial one.

Thin source pool means low competitive floor. A disciplined Hebrew earned-media program captures disproportionate share of model quickly because citation density is lower. Right now, 47% of brands have no GEO strategy. 26% have zero mentions in AI Overviews. The first 12 to 24 months are the maximum arbitrage window.

The window closes. AI systems reinforce their own citation preferences over time. The average domain age of a ChatGPT-cited source is 17 years. Authority compounds. Brands that build citation surface area in 2026 own it for a decade. Brands that wait pay a premium to unseat somebody who moved first.

Where the money actually moves

Across ten major Israeli sectors, a 15 to 25% reallocation toward earned-media-driven GEO over 24 to 36 months implies a national reallocation opportunity in the range of NIS 750 million to NIS 2.4 billion — roughly $215 million to $680 million — before any net-new spending is added.

The sector economics are not uniform.

Banking (NIS 600–900M annual spend, 40–55% AI-exposed). Hebrew AI answers to high-value financial queries default to generic guidance or aggregator content. Major Israeli banks rank well on traditional SEO but are rarely named inside the AI response. Uncontested share of model during a period of record mortgage origination. Indicative GEO reallocation: NIS 40 million to NIS 125 million.

Telecom (NIS 350–500M, 55–70% exposed). AI answers about Israeli cellular plans regularly cite Hebrew reviews containing outdated pricing. Real-time plan updates do not propagate into the retrieval layer, creating a structural lag that quietly erodes acquisition ROI. Indicative GEO reallocation: NIS 30 million to NIS 85 million.

Tech and SaaS export (NIS 5.5–8B, 60–80% exposed). The single most exposed category in the Israeli economy. Over 300 Israeli SaaS companies operate internationally. 42% of B2B buyers open evaluation inside an LLM, and the AI response functions as the shortlist. Indicative GEO reallocation: NIS 500 million to NIS 1.6 billion — the majority of the national opportunity sits here.

Travel, insurance, retail, pharma, automotive, energy, government. Every one of them exposed. Every one of them with an incumbent AI narrative already forming, whether the CMO is watching it or not.

The invisible CAC premium

Run the math on a mid-market Israeli SaaS exporter. 500 new customers a year at $1,500 CAC. Total acquisition spend around $750,000. If 42% of target buyers open evaluation inside an LLM, roughly $315,000 of that spend is exposed to whether the brand is cited in the AI answer during shortlist.

A brand absent from the AI answer is not absent at the click stage. It is absent at the shortlist stage — which is earlier, cheaper to fix, and much more consequential. Industry estimates put the effective CAC premium for AI-invisible B2B tech brands at 15 to 35%. That is a line item on the P&L that most Israeli CFOs cannot see yet.

The narrative dimension

For Israeli corporates with international revenue exposure, the narrative inherited inside an English-language AI answer is not under the company's control unless citation surface area is actively built on high-authority English-language domains. Independent audits of leading AI models have documented uneven framing of Israel-related subject matter in English responses. The default answer is not a neutral answer. It has an author, and the author is whoever built the citation graph first.

What this means, operationally

The work that public relations firms have always sold — earned media placement, executive thought leadership, LinkedIn programs, Wikipedia accuracy, review-platform management, crisis response — is now the primary retrieval layer feeding the AI answer. The work has not changed. The value of the output has.

The correct diagnostic questions for an Israeli CMO or CFO in 2026 are narrow and answerable:

What is our share of model in Hebrew, on our category's top 50 prompts? What is it in English? What earned-media domains are feeding the answer today, and are they ours or a competitor's? What is the sentiment when the model names us? Which platforms cite us, and which do not — only 11% of domains earn citations across both ChatGPT and Perplexity, so this must be tracked platform by platform.

If nobody in the building can answer these questions, the brand is running blind on the surface where a third of consumers and nearly half of B2B buyers now start.

The bottom line

Israel sits at the intersection of three reinforcing tailwinds. A concentrated digital ad economy on platforms that are becoming answer engines. Oligopolistic consumer categories where share of model maps to share of wallet. A Hebrew arbitrage window that is measurable, quantifiable, and closing inside 24 months.

NIS 750 million to NIS 2.4 billion is going to reallocate one way or the other over the next 24 to 36 months. The only variable is who moves first and how much of the compounding surface area they own before the window closes.

AI Communications is a mix of journalism, psychology, and engineering. The audience is now the machine. And the machine is reading Hebrew earned media in Israel right now, whether Israeli boards have added the line item to the budget or not.

---

Ronn Torossian is the founder and chairman of 5W AI Communications, the AI Communications Firm. He is the publisher of Everything-PR and the author of two best-selling editions of For Immediate Release.

Source data: 5W Research, AI and Israeli Marketing (April 2026). Full study at 5wpr.com/research/ai-israeli-marketing.