The after-8pm problem: a composite story from SG bakeries

A note before we start: this is a composite story. The owner 'Kim' and 'Tiong Bahru Bakery & Bites' are not a real customer — they're a stand-in we built from patterns we keep hearing across neighbourhood bakeries in Tiong Bahru, Everton Park, Katong and Joo Chiat. Numbers in this piece are illustrative ranges based on those conversations, not audited results from a single business. We'll publish named case studies once customers are happy to be quoted.

Kim, founder of a small artisanal bakery, always hustled. Her sourdoughs and tarts were locally loved, drawing queues from early morning. But by 8 PM, when the shop closed and her small team went home, a different kind of queue formed: unanswered WhatsApp messages.

Every night, Kim would wake up to 40, sometimes 50, new messages. Customers were asking about custom cake orders, pre-ordering pastries for office events in Raffles Place, or checking if their favourite rye loaf was available. By the time Kim or her manager replied the next morning, many had already taken their business to another bakery down the road in Everton Park, or even a larger chain in town.

This wasn’t just about lost sales; it was about lost connection. Kim prides herself on her personal touch, but she couldn't be online 24/7. Her team was stretched thin enough during operating hours, and adding a night shift for customer service just didn't make financial sense for a small business.

She estimated she was losing at least S$300 to S$500 in potential orders every single night. Over a month, that added up to a significant chunk of revenue. More frustratingly, it was revenue she knew was there, waiting to be claimed, if only she could respond faster.

Then, a friend in the F&B scene up in Katong mentioned she was using an AI tool called Zatty to handle her late-night customer queries. Kim was intrigued, but also skeptical. She sells warmth and handmade goodness; could a robot really do that without sounding… robotic?

After a quick chat with the Zatty team, Kim decided to try it. The setup was surprisingly straightforward. They helped her "teach" Zatty about her menu, her FAQs, her delivery options, and even her special promotions. Within a week, Zatty was deployed.

The change was quiet, almost imperceptible to her customers. After 8 PM, instead of a radio silence, customers received instant, polite replies. Zatty could answer common questions about opening hours, ingredients for allergy sufferers, and even guide them through placing a pre-order via a link to her online store.

Crucially, Zatty didn’t try to upsell or sound like a salesperson. Its job was simply to provide information and facilitate transactions. For more complex queries, like bespoke wedding cake designs or large corporate catering for a firm in Fusionopolis, Zatty would collect the customer's details and flag it for human follow-up the next morning.

Over the following weeks, the pattern owners describe is the same: the pile of unanswered morning messages dwindles, and the order book quietly swells. For a bakery doing S$300–500 in lost late-night orders per night, even recovering a third of that compounds to roughly S$3,000 of additional monthly revenue. That's the order of magnitude we hear, not a guarantee.

'I thought we'd maybe claw back a few hundred bucks,' is how owners typically describe the surprise. The number isn't the headline — the headline is that the revenue was already there, waiting, and the only thing standing between the customer and the order was a reply.

Beyond the revenue, there's the time saved. Owners and managers stop spending the first 1.5 hours of every morning playing catch-up on WhatsApp. That time gets reallocated to menu planning, quality control, and developing new pastries for the weekend rush — effectively an extra half-day of productivity each week.

What Kim learned was that her customers didn’t necessarily crave a human touch for every single interaction. They just wanted quick, accurate information. 1. If it's a simple query, an AI can handle it efficiently. 2. If it's complex, the AI should be smart enough to know its limits and escalate.

‘It’s not about replacing my team,’ Kim insisted. ‘It’s about giving them superpowers. Zatty handles the repetitive stuff, so my people can focus on the creative, truly personal parts of our business. It’s like having an extra pair of hands, but one that never sleeps and never complains when a customer asks for the tenth time if we have gluten-free options.’

For any small business owners still manually sifting through endless DMs after hours, Kim’s advice is simple: look into automating the routine. You might be leaving a lot more money on the table than you think.