Category: Singapore AI Stories

Singapore AI Stories · Retail

The clearest picture of AI retail in Singapore isn’t a concept video — it’s a FairPrice supermarket. What began as a single “Store of Tomorrow” pilot in Punggol Digital District in August 2025 is now rolling out across the island: smart carts that guide you down the aisle, price tags that update themselves, and an AI app that tells staff which shelf needs restocking before anyone notices it’s empty.

Smart carts: checkout in 36 seconds

FairPrice’s smart shopping carts — built with Google Cloud’s Gemini Enterprise Agent Platform — navigate shoppers to items, surface personalised promotions, and let customers scan and pay as they go. According to Computer Weekly, average checkout times at the pilot store dropped from several minutes to 36 seconds, and 48 FairPrice Xtra and Finest outlets are slated to carry the carts by end-2026. A telling detail: the first cart design couldn’t leave the store, shoppers complained, and FairPrice redesigned it — AI rollouts succeed on feedback loops, not launch-day perfection.

Digital price cards: 15,000 man-hours back

Swapping printed shelf labels for AI-generated digital price cards across those 48 outlets is projected to save about 15,000 man-hours and S$138,000 a year, per reporting on the rollout — the cards pull live pricing and promotions and even generate product visuals. Any shop that has ever re-labelled a full aisle on promo-change day will feel that number.

Grocer Genie: an AI assistant for the floor team

The quietest tool may matter most. Grocer Genie gives store teams AI-driven analysis of sales, inventory, and customer satisfaction, plus automated task management — connected to in-store video analytics that detect a stock-out and ping a staff member to replenish. Branch managers describe it replacing eyeballed restocking and manual rostering. That is precisely the pattern of our AI staff assistant concept, and the camera side comes with obligations — our PDPA camera analytics guide covers what Singapore law requires.

The stack behind it

Underneath: cloud data infrastructure, real-time recommendation models for personalised offers, video analytics, and Singpass MyInfo integration — with close to 900,000 members linked so eligible discounts apply automatically at checkout. Each layer maps to something in the AI-assisted shop ecosystem: sensing, deciding, serving.

What smaller retailers can take from this

FairPrice has NTUC-scale budgets, but its playbook is refreshingly copyable: pilot in one store, fix what shoppers push back on, then scale; automate label-and-restock drudgery before anything glamorous; give frontline staff an assistant instead of another dashboard. The SME-priced versions of every one of these — chatbots, e-label systems, forecasting, customer-service AI — exist on government pre-approved lists with grant support; our retail AI software and PSG guide maps them, and our retail services team can scope a one-store pilot of your own.

In this series: AI in Singapore hospitals · Hyundai’s smart factory in Jurong · AI at Tuas mega port.

Singapore AI Stories · Logistics

Singapore handled a record 41.12 million TEU containers in 2024, and by the 2040s it plans to move 65 million a year through a single facility: Tuas Port, on track to become the world’s largest fully automated container terminal. The port that keeps Singapore’s shelves stocked is, increasingly, run by AI — and the ripple effects reach every retailer’s supply chain.

Driverless vehicles, around the clock

Tuas Port, officially opened in September 2022, runs fleets of automated guided vehicles that move containers 24/7 at up to 25 km/h, tracked by RFID transponders embedded in the ground. Each vehicle runs six to eight hours on a 20-minute automated charge, and the electric fleet cuts emissions by roughly half compared with diesel equipment. A central command centre orchestrates automated operations, remote equipment control, and diagnostics.

AI in the water, not just the yard

Beyond the terminal, an AI- and satellite-powered next-generation vessel traffic management system monitors ship movements in real time, optimises berth allocation, and predicts congestion before it forms. PSA International has described using AI across terminal safety and traffic orchestration at Tuas, with workloads spanning hands-free container handling and congestion hotspot detection in Singapore’s waters.

OptETruck: fixing the empty-truck problem

The most relatable story is on the road. Before optimisation, roughly 35% of trucks left the port empty — burning fuel, wages, and hours. PSA’s OptETruck platform, built with HERE Technologies, uses AI to assign jobs and optimise routes in real time. Around 400 trucks — about a fifth of Singapore’s haulage market — have adopted it, with a projected cut of 10 million kilograms of CO2 a year. That’s a classic AI pattern any business will recognise: find the expensive empty runs in your operation and let software fill them.

Why this matters to retailers

Every stockout and every landed-cost dollar in Singapore retail passes through this system. A faster, more predictable port means tighter lead times and less safety stock — which makes AI demand forecasting on your side of the chain even more valuable, because the upstream variability is shrinking. And Singapore’s ambition here is the same one driving the refreshed Retail Industry Digital Plan: automate the repetitive at national scale, keep humans on judgement.

What smaller businesses can take from this

PSA’s wins are really three transferable moves: measure the waste you’ve normalised (35% empty trips was just “how trucking works” until it wasn’t), optimise scheduling with software instead of intuition, and electrify-and-automate the tasks that run around the clock. The shop-scale equivalents — delivery routing, staff rostering, reorder timing — are exactly what we scope in our SME services. New to the jargon (AGV, digital twin, orchestration)? The glossary keeps it plain.

In this series: AI in Singapore hospitals · Hyundai’s smart factory in Jurong · FairPrice’s AI supermarkets.

Singapore AI Stories · Manufacturing

In Jurong Innovation District, a seven-storey building assembles electric cars without a single traditional conveyor belt. The Hyundai Motor Group Innovation Center Singapore (HMGICS) is where the first made-in-Singapore EV — the IONIQ 5 — rolls off cell-based production lines run by humans, AI, and more than 200 robots working together. It’s the most complete picture of AI-driven factory production in Singapore today.

A factory without conveyor belts

Instead of one long line, HMGICS uses flexible production cells: vehicles move between stations on autonomous robots, and each cell can be reconfigured in software. The facility can build up to 30,000 EVs a year — including the IONIQ 5, IONIQ 6, and Kia EV5 — and the whole plant is mirrored by a digital twin, so engineers simulate changes virtually before anything moves on the floor. According to Singapore EDB, the site has reached nearly 70% automation across logistics and manufacturing, with lead times and bottlenecks cut by more than half since the AI-orchestrated robotics ecosystem went live.

Where the AI actually sits

Three layers do the heavy lifting. Vision AI inspects vehicle quality automatically in dedicated inspection cells — including patrols by Spot, the Boston Dynamics quadruped, running AI-enabled defect detection on equipment. Orchestration AI coordinates a fleet of 60 5G-connected autonomous mobile robots across the assembly floor, a deployment supported by IMDA’s 5G Innovation Programme, which reports manual material handling cut by over 50% and zero safety incidents through the implementation. And optimisation AI plans production — the same demand-and-scheduling logic, at industrial scale, that AI demand forecasting gives an F&B kitchen.

Singapore as the world’s testbed

HMGICS was built deliberately small so ideas could be proven fast. It worked: Hyundai Motor Group has transferred around 60% of the innovations piloted in Singapore to its far larger Metaplant in Georgia, USA — meaning production concepts validated in Jurong now shape factories building half a million vehicles a year. That’s the national playbook in miniature: prove it in Singapore, export it globally.

People weren’t removed — they moved up

Roughly half of tasks are performed by robots, but the design keeps humans on supervision, judgement, and exceptions while machines absorb the repetitive, heavy, and hazardous work. It’s the same principle we describe in the AI-assisted shop ecosystem: automation carries the routine load; your team makes the calls that matter.

What smaller businesses can take from this

You don’t need 200 robots to copy the logic. HMGICS wins because it senses everything (data from every station), simulates before committing (digital twin), and automates the repetitive while keeping people on judgement. A shop-scale version is entirely buildable: camera analytics for footfall (done lawfully — see our PDPA camera guide), forecasting before ordering, and automation for the tasks nobody misses. Our step-by-step AI shop guide lays out that path layer by layer.

In this series: AI in Singapore hospitals · AI at Tuas mega port · FairPrice’s AI supermarkets.

Singapore AI Stories · Healthcare

If you want proof that AI has moved from pilot to daily practice in Singapore, look inside its public hospitals. Doctors’ consultation notes now write themselves, referral letters are drafted by a homegrown language model, and chest X-rays are triaged by algorithms — backed by a S$150 million national push into generative AI for public healthcare. Here’s what’s actually running, and what any Singapore business can learn from it.

Note Buddy: the AI scribe in SingHealth clinics

SingHealth’s Note Buddy listens during consultations and turns the conversation into structured clinical notes in real time — in English, Mandarin, Malay, and Tamil. It can tell speakers apart (doctor, patient, caregiver) and formats notes to each specialty’s template. Built on Tandem, a secure GPT platform developed by national healthtech agency Synapxe, it had already supported over 2,100 healthcare workers across more than 16,000 clinical and administrative notes within its early rollout, with generative documentation tools extended across all public healthcare institutions.

The lesson generalises far beyond medicine: the first, safest, highest-ROI AI use case is usually documentation and admin — the same principle behind our guide to choosing your first AI project.

RUSSELL-GPT: NUHS builds its own LLM assistant

The National University Health System developed RUSSELL-GPT, an LLM-based assistant that summarises patient case notes and drafts referral letters — clawing back hours of clinician time per week. NUHS also runs Endeavour AI, which tracks bed availability in real time so patients aren’t stuck waiting for allocation, and Bot-NUHS, which palliative-care nurses use to translate complex care discussions into plain language across Singapore’s languages.

AI that reads X-rays before a radiologist does

Through AimSG, the national radiology AI platform, chest X-ray AI prioritises urgent cases at Geylang Polyclinic, with deployments for tuberculosis screening at the National Centre for Infectious Diseases and bone-fracture detection at Woodlands Health — and a national rollout targeted by end-2026. Patients feel this as faster answers; hospitals feel it as radiologists focused where human judgement matters most.

AI for patients, not just clinicians

Synapxe’s HealthHub AI answers health and admin questions in English, Chinese, Malay, and Tamil, tailored to the user’s age and conditions, while Lab Report Buddy translates blood-test jargon into plain language and flags whether a follow-up is needed. Notice the pattern: the same conversational AI a hospital uses to serve patients at scale is the technology behind retail customer-service chatbots — same engine, different counter.

What smaller businesses can take from this

Healthcare is the most regulated, highest-stakes industry in Singapore — and it still found safe wins by starting with admin, documentation, triage, and multilingual customer communication. If it works under clinical governance, the equivalent works in a shop: notes and rostering, repetitive questions, prioritising what needs a human first. Start where the stakes are low and the drudgery is high — our SME services page shows how that maps to a small business, and the plain-English glossary decodes the terms (LLM, triage AI, speech-to-text) you’ve just read.

Next in the series: inside Hyundai’s AI-run smart factory in Jurong.