Most operators prepare for Passover by building a menu. The callers are not calling about your menu. They are calling to find out what happens in your kitchen, and they will decide whether to keep talking based on how the first question gets answered.
That question is some version of "is this kosher for Passover?" It has a factual answer, it is not a marketing question, and the way most restaurant phones handle it is bad. Someone grabs the phone mid-service, gives a confident-sounding half-answer, and either loses a customer who needed certainty or wins one who is going to be furious later.
The first ninety seconds are screening, not ordering
A normal catering call starts with what the caller wants. A Passover call starts with whether you qualify.
Expect a run of questions before any item is named. Are you certified, and by whom. Is the kitchen changed over or is this a Passover-style menu prepared in a regular kitchen. Do you use kitniyot, which matters to some households and not others. Is the whole menu Passover-appropriate or only part of it. Are there separate meat and dairy preparations. What about the packaging and delivery.
None of that is unreasonable, and all of it is answerable in advance. But it means the phone has to be good at answering questions, not just good at taking orders. An operation whose phone coverage is one busy manager will handle these calls worse than it handles a twenty-piece wing order, because the wing order is muscle memory and this is not.
Write the kashrut sentence once, then never improvise it
Decide exactly what you are, put it in one sentence, and use that sentence everywhere: on the phone, on the website, on the printed menu.
If you hold certification, the sentence names the agency. If you run a Passover-style menu in a kitchen that is not certified and not changed over, the sentence says that in plain words, up front, before the caller invests any more time. There is a real market for Passover-style food among households that are not strictly observant, and being direct about what you are does not shrink it. Being vague and getting caught does.
Whoever answers your phone, person or software, reads that sentence and stops. No paraphrasing, no "I think so," no helpful expansion. If you run a voice agent on the phone, this is the clearest case for configuring an exact-response answer rather than letting a model compose one, and for setting every unmatched kashrut question to escalate to a person immediately. The general approach for religious dietary handling is worth reading alongside voice AI for kosher restaurants.
Lead times come from your kitchen, not from a calendar
The temptation is to publish a round number. Two weeks sounds authoritative. It is also arbitrary, and arbitrary cutoffs are the reason operators end up either turning away business they could have taken or accepting business they cannot cook.
Count backward instead. Start from the first seder. Subtract the prep hours the confirmed orders will consume, which for seder plates and brisket-heavy menus is more per dollar than your usual catering. Subtract the days your specialty suppliers need, which during this window are longer than usual because every kosher kitchen in your area is ordering from the same short list. If you do a full changeover, subtract that too, and be honest that the kitchen produces nothing during it.
Whatever date that lands on is your cutoff. Publish it, put it in your agent's knowledge, and hold it. The cutoff that gets quietly extended for one good customer is the cutoff that fails on the last Thursday when four people call at once.
Then set a second, earlier date: the day you stop taking new items and only take reorders of what is already on the sheet. Menu flexibility is what kills you in the final week, not volume.
Deposits are not optional here
A Wednesday catering cancellation is annoying. A cancelled seder order is a loss you cannot recover, because the demand exists on two specific nights and then evaporates. Nobody is buying your leftover brisket on the ninth day.
So take deposits on anything large. A quarter to half the order value is normal and almost never costs a booking, because the caller planning a seder for eighteen people is not price-shopping at that point, they are securing a spot. Say the deposit amount out loud during the call, along with the date it becomes non-refundable, and send it in writing.
Taking that money over the phone is where operators get sloppy. Card numbers read aloud to a host stand, written on a ticket, and left on a clipboard is a genuine problem regardless of the season. If you are collecting deposits by phone at any volume, read collecting payment over the phone safely and deposits for large catering before this year's rush, not after.
What the phone should be doing during the two weeks before
Volume shifts in a specific way. Order calls spike early, then flatten. Question calls stay high right up to the day, and a rising share of them come from people who already ordered.
Pickup time. Whether the brisket comes hot or cold. Reheating instructions. Whether the delivery driver will arrive before candle lighting, which is a hard deadline and not a preference. Those calls arrive during your service hours and each one is thirty to ninety seconds of a manager's attention at exactly the wrong moment.
That is the honest argument for automating the phone during this window: not that software takes seder orders better than your catering manager does, but that it answers the ninetieth "what time is my pickup" without anyone leaving the line. Confirmed-order questions are the easiest category to hand off and the most expensive category to keep. If you want the sizing math, voice AI for large catering orders covers how these calls actually break down.
Hours, and the day the phone should not be taking orders
Passover has days where observant customers are not calling and days where they are calling constantly. It also has, for many operations, days when your own kitchen is closed or running a reduced menu.
Set those overrides before the week starts rather than during it. An agent that cheerfully takes an order for a day you are closed does more damage than a phone nobody answers, because the customer leaves the call believing they have food coming. The mechanics of doing this properly are in holiday hours overrides for voice AI.
Same for the cutoff. The day after your order deadline, the phone should say the deadline has passed and offer the waitlist, not take the order and route it to a manager who will have to call back and apologize.
The test worth running in March
Call your own restaurant from a number your staff does not recognize and ask, in order: are you kosher for Passover, who certifies you, is the kitchen changed over, and what is your order deadline.
If you get four clean answers in under a minute, your phone is ready. If you get a pause, a transfer, or a confident answer that turns out to be wrong when you check it against your own certification paperwork, you have found the thing to fix, and you have found it while there is still time to fix it.