According to the business press, Bank Hapoalim announced 12 AI projects for 2026 and put Copilot on the desks of thousands of employees, the largest rollout of its kind in Israel. Bank Leumi built an AI center that reports straight to the CEO. The Phoenix is launching LALI, a GenAI avatar for service in general insurance. Bezeq has stopped treating AI as a project and started treating it as infrastructure - a smart router, faults fixed before the customer notices them.
Now go looking for the same announcements from the big names in retail and income-producing real estate. There are almost none. One fashion group recently appointed a chief digital and data officer for the first time in its history, and that is roughly the whole list. The gap between these sectors is the business story of 2026, and if you sit on the management team of a retail chain or a property company, it is worth understanding how it opened and what to do with it.
Why the Banks Went First
The answer is counterintuitive: because of regulation, not in spite of it. The financial regulator's December 2025 guidance requires supervised entities to run structured AI governance, to disclose their use, and to name someone accountable for every application. You would expect that to freeze them. It did the opposite. The rules forced the banks to build the management scaffolding - who approves, who is accountable, which data is permitted - and once that scaffolding stands, running 12 projects at once stops being frightening.
That is the first lesson for the sectors still waiting. Governance is not the brake, it is the accelerator. An organization that knows exactly what it is allowed to do moves faster than one that is nervous about every step.
The Gap Is Not Technological. It Is Managerial
Let me be blunt: the same models serving the bank are available to a retail chain or a property company at exactly the same price. Nobody here has a technology advantage. The difference is the shape of the decision. An AI center reporting to the CEO is first of all a managerial statement - that this belongs on the executive table as a strategic question, not with the head of IT as project number 14.
The national numbers show how rare that statement still is. The State Comptroller's June 2026 report found that only 18% of organizations in Israel hold an AI strategy, 41% operate with no data governance, and 68% of projects never leave pilot. The banks are the exception. The rule is an organization whose employees already use AI privately - Israel leads the world in personal adoption, according to the Anthropic index - while management has not yet decided who owns it.
The pressure is building from the boardroom as well. Across the S&P 500, 83% of companies name AI as a risk factor in their filings, up from 12% in 2023, while only 2.7% of directors hold AI expertise. The question "what are we doing about AI" will reach your table from the board whether you raise it yourself or not. Better to arrive with an answer.
What Is on the Table in Retail and Real Estate
The gap stings because both sectors are sitting on mature use cases. Global industry research puts retail personalization at 10% to 35% of revenue among the companies that implement it, demand forecasting at roughly 50% fewer out-of-stocks and 10% to 20% less inventory, and customer service at savings of around 30%. None of this is speculative. It is running in the world today, with measured returns.
Income-producing real estate looks friendlier still: small head offices, few core systems, and a mountain of documents. Lease agreements to summarize and analyze, tenant service that answers around the clock, tenant underwriting, back-office work that runs itself. A company with 200 leases and three people at head office reading them by hand is exactly where one small system pays for itself within months. It also holds a structural advantage the banks would envy - fewer legacy systems to connect to, fewer committees, and a much shorter distance between a management decision and a working system in the field.
The other side of the coin still applies. Amendment 13 to the Privacy Protection Law, in force since August 2025, covers everyone: fines of up to NIS 3 million and personal liability for officers. A retail loyalty club and a tenant database are precisely the kind of records the law is looking at. Even a company in no hurry to move forward has to put its house in order.
The Window Closes Slowly, Then Fast
Forrester expects 25% of AI budgets to be cut in 2026 for lack of proven ROI. The era of the pilot you run so you can say you have one is over, and what survives is whoever reached a working system. In any sector, the first player to settle its governance and ship one measurable use case sets the standard for everyone else, exactly the way the banks did it to each other.
The one advantage of having waited is that you get to learn from the banks without paying their tuition. Take the proven recipe - governance first, a direct line to the executive, one measurable project with a metric agreed in advance - and run it in a sector where the competition is still asleep. Do that first in a quiet market and you are not closing a gap. You are opening one.
If you are in retail or real estate and want to see what this looks like on a small scale, pick one process, one contract, one report. Send it over and I will send back a working demo before your competitor finishes writing the tender.
