How a Melbourne Restaurateur Running 3 Venues Uses AI to Stay Sane

Composite case study based on real patterns from Australian hospitality operators using AI. Names are illustrative.

Tony runs three Melbourne restaurants: a lunch-focused CBD cafe, a casual Italian in Fitzroy, and a weekend brunch spot in Brunswick. 35 staff across the three venues, two managers, and Tony doing everything else. The hospitality industry runs on thin margins and constant operational pressure. Here’s where AI has actually made a difference.

Menu Costing

Menu engineering: figuring out which dishes make money, which ones don’t, and pricing new dishes correctly: used to take Tony a full day per venue per season. He’d do it with spreadsheets, cursing the whole time. Now he pastes his ingredient costs and portion sizes into ChatGPT along with target food cost percentages, and asks it to calculate gross margin per dish and flag anything below target.

“Here are my ingredients and costs for 8 menu items [paste data]. Calculate the food cost percentage and gross margin for each dish. My target food cost is 28-32%. Flag any items above 35%. For flagged items, suggest either a price increase or portion adjustment to bring them in range.”

Menu costing across three venues: down from a full day to a few hours. He also uses it to model the impact of ingredient price changes: when his chicken supplier increased prices by 12%, he ran the numbers on all chicken dishes in 10 minutes.

Rostering

Tony uses Deputy for rostering: the Australian-built workforce management platform that already has AI-assisted scheduling built in. It factors in forecasted covers (based on historical data), staff availability, and Fair Work Act requirements around minimum breaks and penalty rates. Weekly rostering time: from 3 hours to under 45 minutes across all three venues.

Google Review Responses

Three venues means a lot of reviews. Tony was responding to maybe 20% of them: the ones he had time for. Now he batches review responses weekly. He copies the reviews into ChatGPT, asks for personalised responses that address specific feedback, and pastes the approved responses into Google. Every review now gets a response within a week.

For negative reviews, he writes his own responses: he doesn’t trust AI to handle those with the right nuance. But he’ll use ChatGPT to draft a first version and then rewrite it significantly.

Supplier Communications

Negotiating with suppliers, chasing credits for short deliveries, writing formal complaints about quality issues: all tedious but important. Tony uses ChatGPT to draft these communications. “A firm but professional email about a supplier consistently delivering short on weight”: the AI drafts it, he reviews and sends. He estimates this saves about 2 hours a week across the three businesses.

Staff Communications

Policy updates, safety reminders, new procedure rollouts. Tony uses ChatGPT to draft these so they’re clear, consistent, and professional rather than a rushed email fired off between service. He’s also used it to write job ads and draft interview questions for new hires.

What It Costs Across Three Venues

• ChatGPT Plus: ~AU$28/month
• Deputy (already in use, 35 staff): ~AU$350/month
• New AI cost: just AU$28/month (Deputy was already there)

Tony’s view: “The ROI conversation doesn’t even make sense. The question is just: why wasn’t I doing this sooner?”

What He Hasn’t Automated

Anything customer-facing in real time. The hospitality experience is human: the greeting, the service, the vibe. Tony is emphatic that AI doesn’t touch the floor. It handles the desk work so he can spend more time on the floor.

📖 Related: Full AI guide for restaurants | AI for cafe owners

Related Reading

• AI for Australian Restaurants
• AI for Cafe Owners
• Deputy: The Sydney Startup Behind Australian Hospitality Rostering
• AI for Google Reviews

📊 Compare AI tools side by side | 💼 Free resources & AI prompt packs

📋 Browse all Australian small business AI case studies →