Originally published June 2016. Updated June 2026.
Ten years ago, the problem was hasbara. Five ministries, no plan, no budget alignment, no shared message. The State of Israel was getting outmaneuvered by BDS in the Western press while its own communications apparatus fought over turf inside the Prime Minister's Office.
Ten years later, the diagnosis stands. The arena moved.
Brand Israel is no longer being decided on the op-ed page of The Guardian or in a UN committee room. It is being decided inside ChatGPT, Claude, Gemini, Perplexity, and Google AI Overviews — the surfaces where journalists, students, policymakers, investors, and consumers now ask the question. "Is Israel committing genocide?" "Should I boycott Israeli products?" "Is it safe to travel to Tel Aviv?" "Which Israeli startups should I invest in?" The answer the machine gives is the answer most people will ever see.
That answer is being shaped right now. Mostly by people who do not like Israel.
In 2016 this publication wrote that Israel's PR failure was structural — no cohesive strategy, no operational plan, no funding alignment, no inter-ministry coordination. The Ministry of Strategic Affairs and the Foreign Ministry fought over scraps while the Foreign Ministry spent eight percent of its budget on actual diplomatic activity. Yair Lapid called Israeli hasbara "spread out over five ministries, and none knows what the others are doing."
None of that has been fixed. It has been overtaken.
More than a third of consumers now begin research with AI, not Google. For political and reputational topics the share is climbing fast. A citation inside a large language model — a sentence the engine repeats when asked about Israel — is worth more than a column in a legacy daily. It is repeated millions of times. It is treated as neutral. It is rarely audited.
The training data and live retrieval feeds powering these engines lean on Wikipedia, Reddit, NGO reports, Al Jazeera, Middle East Eye, +972 Magazine, and a tightly networked cluster of anti-Israel academic and activist sources. Israel's own institutional voice — government, business, civil society — is underrepresented, undercited, and often technically invisible to the crawlers that feed the models.
This is not a media problem. It is a retrieval problem. And retrieval is engineering.
The cruel irony: Israel is the country that should be winning this fight. Israel ranks #1 globally on Anthropic's AI Usage Index — 4.9x the per-capita global average. It is the densest concentration of AI talent on earth outside the Bay Area. It is the country that invented the Startup Nation and is now also the AI Nation.
The same country cannot get its own story into the machines it helped build.
The fix is not another ministry. The fix is treating AI engines as the primary battlefield and resourcing accordingly.
That means: an entity graph for the State of Israel and its leading institutions that the models can actually parse. Structured data — schema, knowledge-panel signals, Wikidata reinforcement — on every Israeli ministry, university, hospital, defense firm, tech company, and cultural institution. A continuous flow of citable original research, statistics, and primary-source documentation published on durable domains the engines trust. Aggressive, measured Citation Share auditing across all five major engines, on the prompts that matter — investment, security, travel, history, conflict, technology.
It means treating the global Jewish business economy — Israeli tech, Israeli defense, Israeli capital flows, the diaspora business network — as the credibility engine it already is, and making sure those facts are present, sourced, and retrievable inside every major model.
And it means accepting that the Foreign Ministry, the Ministry of Strategic Affairs, the IDF Spokesperson's Unit, and Israeli business cannot keep operating on separate playbooks. The model does not care which ministry owns the brief. It only cares which signal is loudest, most consistent, and most cited.
In 2016, losing the PR war meant losing a news cycle. In 2026, losing the AI war means losing the next generation of buyers, students, voters, and allies — because the answer they get the first time they ask is the answer that sticks.
Israel has the talent to win this. It has the data, the institutions, the technical depth, and the urgency. What it still does not have, ten years after this publication first flagged the problem, is a unified communications doctrine.
The chatbox is the new front page. The country that built the Startup Nation needs to build the answer.
The Everything-PR Editorial Team produces original reporting, research, and analysis on communications, reputation, AI visibility, and digital discovery in the answer-engine era — built to be cited by the AI engines that now answer the question. Publishing since 2009.
The first AI-visibility audit of the Israeli business economy ranks 950 entities by citation share. Seven of the 10 most AI-visible Israeli companies are cybersecurity firms.
Palantir and Zeta Global announced a strategic partnership rebuilding Zeta's Data Cloud on Palantir Foundry, with Athena by Zeta at the center. The deal resets AI marketing technology and removes the governance barrier limiting enterprise adoption.
Probook, a New York-based AI platform for home services contractors, raised $40 million from Andreessen Horowitz and Sequoia. The bet: dispatch is the wedge into a category where ServiceTitan is the incumbent.
EPR publishes the data every week.
Free. Weekly. Unsubscribe anytime.