Most restaurants and food brands have a nutrition spreadsheet behind the scenes that customers never see. The brave ones link to a 30-page PDF. Neither approach answers the actual question a customer has, which is usually something like 'which of your bowls fits a 600 calorie lunch with 40g of protein'. That answer requires reading the data the way the customer would, not the way the spreadsheet is laid out.

SleekAI reads your menu_items or products custom post type, the nutrition postmeta with per-serving calories, protein, carbs, fibre, fat, saturated fat, sugar, sodium, and any diet tags like vegetarian, vegan, high-protein, keto. The bot can answer 'which bowls are under 600 calories with at least 40g of protein' or 'is the chicken wrap higher in protein than the salmon bowl' without anyone consulting the PDF.

For brands that update recipes seasonally, edits to the nutrition fields show up in the bot's answers immediately. Display conditions scope the widget to the menu, product pages, and any meal-plan area. Logs capture the dish, the constraint asked about (calories, protein, sugar), the model, and the tokens, which gives the kitchen a useful read on what customers actually care about.

Nutrition data sits in a strange place on most food sites. Compliance teams require it to exist, and it usually does as a PDF deep in the footer or a long table on a product page. Customers see it because they were forced to or because they were already determined to find it.

Almost nobody discovers a dish because of the nutrition data, even though the nutrition data is exactly what would help them choose. A chatbot flips the dynamic. The customer arrives with a goal in their head: a lunch under 600 calories, a high-protein vegan option, a low-sugar bar.

The bot turns the goal into a query against your real per-serving figures and returns dishes that fit, with the numbers right there for verification. That is closer to how a knowledgeable friend would help someone order than to how a compliance PDF works. The kitchen and product team get a feedback loop they never had before.

They learn which macros customers care about, which constraints are hardest to satisfy on the current menu, and which combinations of restrictions (vegan plus high protein, keto plus low sodium) attract the most queries. A few months of that data is a product brief in disguise. New menu items can be designed to fill the gaps the chatbot has been quietly mapping.

The chatbot is not a calorie counter, it is the interface between what the kitchen documents and what the customer actually wants to choose.