SMEs want AI too, but not the kind Big Tech is selling

With additional reporting by Scott Shuey

Cakeinspiration has made a name for itself in Singapore by getting elaborate cakes into the hands of visiting celebrities, including American singer Jason Derulo and K-pop girl group Blackpink.

But for Kai Yang Chan, who runs the family bakery business with his wife, the customers he was more worried about were the ones waiting for a reply on WhatsApp.

The shop’s marketing brings in a steady stream of enquiries – from people ordering a standard cake to customers with highly specific designs in mind.

Chan’s sales team used to answer each one manually, but messages would pile up, as details had to be collected one by one.

Now, an AI sales agent replies to routine queries, collects the basics, and separates standard orders from custom ones before passing the lead to a salesperson.

Chan says sales have improved by “at least”10%.

“With the AI helping them, I guess [the employees] feel less stress,” he tells Tech in Asia.

Chan is now looking to add a delivery agent that can consolidate addresses, timings, and other order details for Cakeinspiration’s drivers.

Another business leveraging AI similarly is Happy Fish Swim School, which runs about eight AI bots across its 21 branches in Singapore and Malaysia. A coach bot answers instructors’ questions and finds substitute coaches, while operations, HR, task, customer relationship management (CRM), and inventory bots help train staff, manage payroll, track customer waiting lists, and monitor stock.

Jian Yong Tan, the swim school’s founder, says the AI systems have given managers visibility they had begun to lose as the business grew.

But Cakeinspiration and Happy Fish are a minority when it comes to SMEs adopting AI agents. Even in more advanced markets, SMEs have been slow to pick up AI.

Across the OECD, an international network of 38 high-income countries, 40% of large firms use AI, compared with just 11.9% of small firms. And even among SMEs that use generative AI, only 29% have applied it to core business activities.

Mom and pop AI

Stripped down to the essentials

Keeping data out of the black box

The luxury of GPUs