Pairings and Upsell

A guest orders a biryani. Do they also want a raita, a cold drink, or a dessert? DineomAI asks for you, right inside the chat. There are two ways it sells more: pairings you set by hand, and an AI upsell that suggests on its own. You manage pairings at /admin/pairings.

Set up a manual pairing

A pairing is a simple rule: when a guest orders X, suggest Y. You pick the trigger item and the suggested item, and DineomAI offers the suggestion at the right moment in the chat.

Think about what goes together on your menu. A spicy main pairs well with a cooling drink. A coffee pairs with a slice of cake. You know your menu better than anyone, so start with the matches that feel obvious to your regulars.

Read the 30-day stats

Each pairing carries its own numbers, measured over the last 30 days, so you can see what's working:

• Trigger orders: how often guests ordered the trigger item.
• Accepted: how many guests took the suggestion.
• Conversion rate: the share of triggers that turned into an accepted suggestion.
• Revenue: the money those accepted suggestions brought in.

If a pairing converts well, keep it. If it barely moves, swap the suggested item for something else and watch the next 30 days.

Accept an auto-pairing

DineomAI watches what guests order together often and proposes auto-pairings from that pattern. When you see a suggestion that fits, add it with one tap. It's a fast way to find pairings you might not have spotted, since the data sees combinations across every table.

Turn on AI upsell

AI upsell is the second lever. Instead of a fixed rule, the AI generates a pairing suggestion and shows it to a share of your guests. You set it up during onboarding, where you choose three things:

• On or off: whether the AI upsell runs at all.
• Percent of guests: what share of guests see an AI suggestion.
• Monthly cost cap: a ceiling on what the feature can spend in a month.

The cost cap keeps the bill predictable. Once the month's spend hits the cap, the upsell pauses until the next month.

What to expect

Across a month, upselling tends to lift the average ticket by about 5 to 10 percent. That figure is approximate. It varies by restaurant and by menu, and your own number depends on your prices, your guests, and the pairings you pick.

Here's a rough example. Say a 40-seat restaurant in Bangalore runs an average ticket of 600 rupees across 1,500 orders in a month. A 7 percent lift adds about 42 rupees per order, or roughly 63,000 rupees that month. Your real result will differ, so watch your own analytics rather than this estimate.

FAQ

What's the difference between a manual pairing and AI upsell?

A manual pairing is a fixed rule you set: when a guest orders X, suggest Y. AI upsell is generated by the AI and shown to a share of guests, with a percent and a monthly cost cap you control. You can run both at once.

How do I know if a pairing is actually working?

Open /admin/pairings and read the 30-day stats for that pairing. Look at the conversion rate and the revenue. If guests rarely accept it, change the suggested item and check again after a few weeks.

How much will upselling add to my sales?

Across a month, it tends to lift the average ticket by about 5 to 10 percent. This is approximate and varies by restaurant and menu. Track your own analytics to see your real number.

What to do next

Add one or two manual pairings for your best-selling dishes, then accept any auto-pairings that fit. Pair this with your offers and discounts to give guests a reason to add more.