Cooking for 200 people isn’t just a bigger version of cooking for 10. It’s a logistics problem, a math problem, and a dietary-restriction minefield — all happening at the same time, usually under serious time pressure. Edy Massih, the owner of Edy’s Grocer, a Lebanese-inspired grocery and catering operation, figured out that Google Gemini could handle the parts of large-scale catering that used to eat up hours of manual work. Google’s own breakdown of how Massih uses Gemini reveals three specific workflows that are genuinely worth paying attention to — not because they’re flashy, but because they’re practical in a way that most AI use cases aren’t.
The Problem With Scaling Food for a Crowd
Here’s the thing about catering at scale: the math sounds simple until you’re actually doing it. Tripling a recipe doesn’t always mean tripling every ingredient. Spices don’t scale linearly. Baking times change. Batch cooking introduces variables that home recipes were never designed to account for.
For a small business owner running a grocer and a catering operation simultaneously, that kind of mental overhead compounds fast. Massih isn’t running a hotel kitchen with a full brigade of chefs. He’s working lean, which means every hour spent manually converting recipes or cross-referencing ingredient lists is an hour not spent on something else.
That’s the context here. This isn’t a story about a tech company deploying AI at enterprise scale. It’s about one person using a general-purpose AI tool to solve very specific, very real operational problems. And frankly, those stories are often more instructive than the big-budget deployments.
Three Ways Gemini Actually Gets Used
1. Scaling Recipes From Dinner Party to Event Catering
The first use case is the most straightforward: Massih feeds Gemini a recipe designed for, say, 6 people, and asks it to scale up to 200 servings. But it goes further than simple multiplication. He’s reportedly using Gemini to flag where scaling ratios need adjustment — where a spice that works at small quantities becomes overpowering at volume, or where a sauce that emulsifies easily in a small batch needs a different technique when made in bulk.
This is actually a harder problem than it sounds, and it’s one that professional catering software has tried to solve for years. Tools like Caterease and Total Party Planner offer recipe scaling features, but they’re rigid — they apply multipliers without any culinary judgment. Gemini, being a large language model trained on an enormous corpus that presumably includes culinary knowledge, can apply a layer of reasoning that a spreadsheet formula simply can’t.
The practical output is a modified recipe with adjusted quantities and, where relevant, notes on technique changes. That’s something a seasoned chef carries in their head. For a small operation that doesn’t have that institutional knowledge baked in, it’s genuinely useful.
2. Automating Grocery Orders
Once the scaled recipe exists, the next step is procurement. Massih uses Gemini to generate grocery orders based on the adjusted ingredient lists — cross-referencing what’s already in stock and producing a consolidated order list for suppliers.
This is where the time savings get concrete. Manually building a purchase order for a 200-person event means going line by line through multiple recipes, checking inventory, calculating quantities with buffer for waste, and formatting everything for supplier submission. For a busy event week, that could easily run two or three hours of administrative work.
With Gemini handling the aggregation, that process compresses significantly. The AI isn’t just copying ingredient lists — it’s reportedly consolidating across multiple dishes, identifying overlapping ingredients, and producing a single unified order. That kind of multi-document synthesis is exactly what large language models are well-suited for, and it’s a legitimate time-saver in a context where time is genuinely scarce.
3. Adapting Menus for Dietary Restrictions
The third use case might be the most operationally complex. Modern catering means fielding guest lists with a wide range of dietary requirements — gluten-free, vegan, nut allergies, halal, kosher, low-FODMAP, and combinations thereof. Managing substitutions across a full menu, while keeping dishes coherent and appealing, is a real cognitive load.
Massih uses Gemini to take a base menu and generate modified versions for specific dietary profiles. The AI suggests ingredient swaps, flags potential cross-contamination issues, and produces alternative preparations that maintain the spirit of the original dish.
This is an area where I’d expect some limitations — Gemini isn’t a registered dietitian, and for guests with serious allergies, any AI-generated substitution list would need human verification before it becomes the actual menu. But as a first-draft tool that handles the bulk of the substitution work and surfaces the edge cases for human review, it fits neatly into a realistic workflow.
What This Actually Tells Us About AI Adoption
There’s a pattern worth noticing here. The AI applications getting real traction in small business aren’t the dramatic, autonomous agent scenarios that dominate conference keynotes. They’re tools that compress the time spent on high-repetition cognitive work — drafting, calculating, reformatting, cross-referencing.
Massih’s use of Gemini fits that pattern exactly. None of the three workflows involve AI making autonomous decisions. A human is still designing the menu, approving the grocery order, and signing off on dietary accommodations. Gemini is handling the labor-intensive middle steps that don’t require creativity or judgment — just accuracy and speed.
That’s a healthy division of labor, and it’s probably a more durable model for AI adoption than the fully-automated alternatives. It also points to something interesting about where Gemini is finding its audience. Google has been pushing Gemini hard across its Workspace suite, and the integrations are maturing. For small business owners already living in Google Docs and Sheets, the friction to start using Gemini for this kind of task is low.
Compare that to the developer-facing applications we covered in our look at what developers built with Gemini 3.8 Flash — sophisticated API integrations, custom tooling, complex prompting chains. Massih’s use case is on the opposite end of that spectrum, and both are valid. The breadth of that range is actually what makes Gemini an interesting product to watch right now.
It’s also worth putting this alongside other small-business AI success stories. We’ve seen Proaction boost sales 60% using OpenAI Codex in a completely different industry context. The throughline is the same: identify the repetitive cognitive work, hand it to the AI, keep humans in the loop for decisions that actually matter.
Key Takeaways
- Recipe scaling with culinary judgment: Gemini goes beyond simple multiplication, applying reasoning about how ingredient ratios and techniques need to shift at volume.
- Consolidated grocery ordering: The AI aggregates ingredient lists across multiple scaled recipes, cross-references inventory, and produces unified supplier orders — compressing hours of admin work.
- Dietary restriction management: Gemini generates modified menu versions for specific dietary profiles, flagging substitutions and potential issues for human review.
- Human oversight remains central: None of these workflows are fully automated. Gemini handles the labor-intensive drafting; Massih makes the final calls.
- Low barrier to entry: For small business owners already in Google’s tools, Gemini integration requires minimal technical setup.
Frequently Asked Questions
What is Edy’s Grocer using Gemini for?
Edy’s Grocer uses Google Gemini to scale recipes from small servings to 200+ guests, generate automated grocery orders based on scaled ingredient lists, and adapt menus for guests with dietary restrictions. The goal is to reduce the manual administrative time involved in large-scale catering operations.
Does Gemini replace professional catering software for this use case?
Not exactly — dedicated catering platforms like Caterease offer features Gemini doesn’t, including client management, invoicing, and event scheduling. But for the specific tasks of recipe scaling, ingredient aggregation, and menu adaptation, Gemini’s language understanding gives it an edge over the rigid multiplier-based scaling in most traditional catering software.
How reliable is AI-generated dietary substitution advice?
For guests with serious allergies or medical dietary requirements, any AI-generated substitution list should be treated as a starting draft and verified by a human with relevant knowledge before it becomes an actual menu. Gemini can surface most common substitutions accurately, but it’s not a substitute for professional dietary review in high-stakes situations.
Is this kind of AI use accessible to other small food businesses?
Yes, and that’s arguably the most interesting part of this story. Massih’s workflows don’t require API access, custom development, or technical expertise. If you have a Gemini subscription and can write a clear prompt, you can replicate most of what he’s doing. The barrier here is knowing what to ask for, not building anything.
As Gemini continues to get embedded deeper into Google’s productivity tools, expect more small business use cases like this to surface — specific, unglamorous, and quietly effective. The question isn’t whether AI can handle a recipe for 200 people. It’s how many other time-consuming back-office tasks in food, retail, and services are sitting there waiting for someone to try the same approach.
