Israel’s ₪49 AI Receptionist Boom
There is a drawer in my office with seven SIM cards in it.
I bought them this summer for an unglamorous reason: I build phone-answering software, and I needed to know whether the standard Israeli call-forwarding code — the little *004* incantation that tells your carrier “if I don’t pick up, send the call there” — actually works on real Israeli networks. The answer, after seven SIMs and four carriers, was sobering: it worked out of the box on exactly one of the four carriers. On the others, the code that every international tutorial promises you either failed silently or was blocked somewhere in the network. Nobody advertises this. You find out after you’ve paid.
I’m telling you about the drawer because Israel is suddenly full of ads for “AI secretaries,” and most business owners I talk to cannot tell the products apart — or why one costs ₪49 a month and another ₪549. Having spent the last year inside this market (full disclosure: I built one of these services), let me map it honestly.
Why this category exists at all
Start with the problem, which is bigger than most owners admit. A 2024 study by 411 Locals monitored 85 small businesses across 58 industries and found that only 37.8% of incoming calls were answered by a live person. Roughly another third went to voicemail — where, according to data popularized by CallRail and BrightLocal, some 85% of callers simply hang up without leaving a message. Most of those callers do not try again; they dial the next result on Google.
For an Israeli clinic, law office, or renovation contractor, that arithmetic is brutal. The plumber is under a sink at exactly the moment his next month of work is calling. The clinic’s receptionist goes home at 16:00; anxious patients do not.
Until recently, your options were a human answering service (hundreds of shekels a month, script-bound), the carrier’s voicemail (see above: people hang up), or losing the call. What changed in 2025–2026 is that speech AI in Hebrew became good enough — and cheap enough — to answer a phone call convincingly. And so, within months, an entire product category materialized. The trouble is that three genuinely different products are all being marketed with the same words: mazkira AI, “virtual secretary,” “AI receptionist.”
The three tiers, decoded
Tier 1: the missed-call secretary (₪49–79 a month). These services answer only the calls you didn’t take. The caller hears a natural Hebrew voice, states their business, and you get a transcript and summary, typically by WhatsApp, seconds later. Two Israeli services define this tier: Maya (hellomaya.co.il), which advertises answering, transcription and WhatsApp summaries “from ₪49 a month,” and Onim Li (onimli.com) — the one I built — at ₪49 a month for individuals and ₪79 for businesses. This tier doesn’t schedule appointments and doesn’t replace anyone; it converts the black hole of voicemail into a readable message. Its economics work because the AI only speaks for the minute or two after a missed call.
Tier 2: the full receptionist (₪179–549 a month). Here the AI answers every call, first ring, all day — it qualifies the caller, answers common questions, and books appointments into your calendar. Genie (getgenie.co.il) is the visible Israeli player in this tier; its site quotes two plans, basic at ₪179 a month and growth at ₪549, with no commitment and a seven-day trial. The price jump over Tier 1 is honest: real-time answering of live calls means integrations, longer conversations, and much more AI compute per customer.
Tier 3: the custom voice agent (project pricing, thousands of shekels). Above the subscriptions sits bespoke work: a voice agent wired into your CRM, your booking system, your price list and your escalation rules. This is not a product you subscribe to but a system that boutique automation studios such as Achiya Automation design and build per business — which is why it is priced like a project, not a phone plan. It makes sense at call volumes where a missed booking costs more than the build.
The confusion — and some of the marketing — comes from blurring these tiers. A ₪49 service demoed as if it were a full receptionist will disappoint you by Thursday. A ₪549 receptionist sold to a one-person studio that misses four calls a week is overkill by an order of magnitude.
What the price never tells you
Three things I learned building in this market, none of which appear on any pricing page.
First, the SIM drawer: activation depends on your carrier, not on the product. Every service in Tiers 1 and 2 relies on call forwarding, and Israeli carriers implement it inconsistently. Before you subscribe to anything, ask the vendor directly: “Has this been tested on my carrier?” A vendor who answers that question in detail has done the work; one who says “it’s a standard code” has not opened the drawer.
Second, Shabbat is a product decision. An AI that answers your phone on Friday night is, for a large share of Israeli businesses, not a feature but a liability. When we built our Shabbat mode we chose to make it real — the system goes fully silent, not merely polite. Whatever you buy, ask what happens at sundown on Friday. The answer tells you whether the product was designed in Israel or merely translated into Hebrew.
Third, memory is the difference between a gimmick and a service. The first time a returning customer is greeted as a stranger, the magic dies. Ask any vendor: does the system remember that this number called last Tuesday about a quote? Persistent caller memory is technically expensive and commercially invisible — which is exactly why its presence signals a serious product.
The honest buying advice
Count your missed calls for one week — your phone’s call log already knows. If you’re missing a handful of calls a week, a ₪49–79 missed-call secretary pays for itself with a single rescued lead, and you’ll feel the difference in a day. If your front desk is drowning and callers wait, that’s Tier 2 money — ₪179 and up — and worth it. If your calls follow complex rules and feed a CRM, you’re shopping for a Tier 3 project, and no subscription will save you.
But pick a tier before you pick a brand. In a market this young, the vendors’ words will keep blurring; your call log doesn’t.
Which call would you still refuse to hand to a machine — the angry customer, the new lead asking “how much,” or your biggest client checking in? That answer, more than any pricing page, tells you which tier you actually need.
