AI re-engineered our Zomato pricing when parity dropped

On April 23, 2026, Zomato quietly dropped the clause that let it fine restaurants up to 3× the price differential for charging less at the table than on the app. NRAI president Sagar Daryani said what every owner had been thinking for two years: "It's our product and should be our pricing." Ashish, who runs a 48-item multi-cuisine menu out of a single Pune outlet, read the news at 11am between lunch covers. By 11pm the same night, his menu had a new dine-in price on 19 dishes — computed, reviewed, and pushed live, not by a consultant, but by an AI pass he triggered with one WhatsApp message.

Most owners will take two to three weeks to get to the same place, repricing dish by dish, on gut feel, usually only after a customer complains that "Zomato pe sasta hai." This is the how-to for doing it in one pass, the day the news breaks.

What Zomato's price parity clause removal actually means for a single-outlet dine-in

For two years, the parity clause was the invisible ceiling on every dine-in pricing decision an Indian restaurant made. If your Zomato-listed price for Butter Chicken was ₹380 and your dine-in menu said ₹340, Zomato could — and did — fine you up to 3× that ₹40 gap. So owners either kept prices identical everywhere, or quietly ate the aggregator's 18-28% effective commission on every delivery order without ever building a real reason for a customer to choose the table over the app.

That ceiling is gone now, confirmed and operator-facing, not a rumour or a workaround. A dine-in price of ₹280 on a dish listed at ₹340 on Zomato is fully legal, fine-free, as of April 23. What that opens up isn't "charge more on delivery" — it's the ability to price each item on its actual economics per channel, for the first time since aggregators went mainstream in India. The removal is the trigger event. The work is figuring out, item by item, where the gap should be ₹0, where it should be ₹60, and where it shouldn't move at all.

The AI pricing model: how it computes optimal dine-in vs aggregator differential

This is the part that used to take a consultant a week of spreadsheet work and now takes an AI pass one prompt to run, provided you feed it the right four inputs:

1. Contribution margin per item — price minus food cost minus packaging, per dish.
2. Effective aggregator commission — not the headline 18% or 22% Zomato quotes, but what you actually net after commission, GST-on-commission, and ad-spend deductions on that specific dish over the last 30 days.
3. Order mix — what percentage of each dish's volume comes through Zomato/Swiggy versus walk-in dine-in.
4. Dine-in discount floor — the minimum gap you're comfortable holding before it looks like a fire sale to a regular who eats at both.

Feed those four columns into Claude or ChatGPT alongside your menu, and it computes a per-item recommendation: hold the price flat, widen the dine-in discount, or actually raise the aggregator-side price to cover the commission bite you were quietly absorbing. The output isn't a single "discount %" — it's a row-by-row table, because a high-margin dine-in main and a delivery-only combo have nothing in common economically, even if they sit on the same page of the PDF menu.

The re-engineering run: what changed on 48 items in one pass

Ashish's 48-item menu run is illustrative — his exact numbers, run once, on one outlet — not a universal formula. Here's the shape of what moved:

• High-margin dine-in mains that had been priced flat across channels for a year — his Mutton Rogan Josh, Paneer Lababdar, and two thali sets — got an +8-12% dine-in premium recommendation on the aggregator side, since 70%+ of their volume was already walk-in and the dine-in guest wasn't price-shopping against the app.
• Delivery-unfriendly items — a fish curry that arrives cold, a dessert that doesn't travel, a soup — got the widest dine-in advantage. These items lose almost nothing to the aggregator channel because almost nobody orders them for delivery anyway; the model recommended holding dine-in price low and low-key, and quietly raising the Zomato-listed price to cover the commission on the rare delivery order.
• Commodity items — plain rice, a basic dal, a lassi — didn't move at all. Delivery customers on these are the most price-sensitive segment on the menu, and the model correctly flagged that any dine-in/aggregator gap here risks looking punitive rather than strategic.

Net result on the 19 items that changed: an estimated ₹14,000-18,000/month margin uplift for Ashish's outlet, by his own back-of-envelope math after the pass — a number specific to his order mix, not a promise for every 48-item menu.

How to trigger the same AI repricing for your menu (the exact prompt + data inputs)

The trigger is one message. Ashish's read: "Zomato parity clause hata — repricing run karo, dine-in premium recommend karo." That's the whole ask.

Behind that one line, here's what actually needs to happen, whether you're typing it into a WhatsApp-native AI back-office or pasting into Claude/GPT yourself:

1. Export the four inputs — menu with food cost, 30-day order mix by channel, your actual (not headline) aggregator commission over the same period, and your comfort floor for the dine-in gap.
2. Ask the model directly: "Here's my menu, my food cost, my order mix by channel (dine-in vs Zomato vs Swiggy), and my effective commission per platform. Recommend a per-item dine-in price and aggregator price. Flag which items shouldn't move. Estimate the monthly margin delta."
3. Review by category, not item-by-item — approve the mains as a batch, the delivery-only items as a batch, the commodity items as a batch (usually "no change"). Push the approved batch live across Petpooja, Zomato, Swiggy, and Google Business in one shot.

The whole loop — export, prompt, review, push — is a same-day job, not a two-week project.

What the margin looks like 30 days after AI-driven repricing

The 30-day view isn't a straight line up. Dine-in and delivery respond differently to the same gap, and the risk worth watching is channel cannibalization — a regular dine-in guest who notices the Zomato price is now visibly higher and switches to ordering in instead of walking over, which erodes the very premium you just built. AI's job in week two through four is to watch the actual dine-in vs delivery mix weekly against the pre-repricing baseline and flag if any single item's dine-in share drops faster than 5-10% — a signal the gap widened too far for that dish specifically.

The counter-argument worth addressing directly: "Zomato will find another way to penalize me for this." The parity clause removal is confirmed, dated, and operator-facing — this isn't a policy that quietly reverses itself. The real risk is customer perception, not aggregator retaliation, which is why keeping the pricing consistent and intentional across every platform — not randomly different — matters more than the size of the gap itself.

So what now

Running this pass by hand across 48 items, four data inputs, and three channels is exactly the kind of one-time-per-news-event pricing strategy a back-office team should own for you — which is the same job SideKyk's ops agent does the moment a Zomato or Swiggy policy shifts: one Hinglish message in your WhatsApp, and the repricing recommendation, category review, and same-shot push across every platform is done before your next billing rush. Join the restaurants waitlist at sidekyk.ai/restaurants — we'll WhatsApp you the moment your slot opens.