This is Week 1 of the Singapore Food Delivery AI Visibility Index — measuring which apps AI engines name for food delivery and super-app utility in Singapore.

The run draws from the AI Visibility Week 1 experiment (manifest 2026-06-29) across ChatGPT, Google AI Mode, Perplexity, and Gemini.

Headline finding

Grab wins overall. foodpanda wins on price. Every head-to-head answer says install both.

AI has learned a stable attribute split — not a single champion. Grab owns quality, speed, selection, and ecosystem. foodpanda owns value. Direct comparisons refuse to pick a winner.

Locked prompt kit (core index)

ID Prompt theme Runs
P69 Best food delivery app (SG) 3
P70 Fastest delivery (SG) 3
P71 Cheapest food delivery (SG) 3
P72 Most useful super app (SG) 3
P73 Grab or foodpanda better (SG) 3
P74 Best restaurant selection (SG) 3

Brand prompts (P75–P78) cover Grab reliability, foodpanda service quality, Deliveroo’s operating status, and Grab vs other SEA apps. SEA prompts (P79–P81) ask for the best food delivery app in Southeast Asia from SG, MY, and ID contexts.

Visibility matrix — attribute split

Prompt theme Consensus winner Notes
Best food delivery app GrabFood Often hedged: “depends / use both”
Fastest delivery GrabFood ETA still framed as location-dependent
Cheapest delivery foodpanda ChatGPT most reluctant to crown a single cheapest
Restaurant selection GrabFood Post-Deliveroo breadth narrative
Super app Grab Unanimous
Grab vs foodpanda Neither — install both 12/12 runs
Grab reliable? Yes, with surge/delay caveats Brand prompt
foodpanda service quality Negative public sentiment Delays, CS, refunds
Deliveroo still operating? No (exited) Correct when asked directly

What Week 1 shows

The duopoly is the product. After Deliveroo’s exit, AI answers treat Singapore food delivery as Grab vs foodpanda. Niche apps (WhyQ for hawkers, Oddle for premium) appear mainly in longer ChatGPT/Gemini answers — never as default winners.

“Install both” is stronger than any brand win. On P73, every model and every run produces attribute tables and a dual-app recommendation. For marketers, that means AI visibility is won at the attribute level (speed, price, selection), not only at “best app.”

foodpanda’s paradox is real. The same experiment week crowns foodpanda for cheapest delivery and documents poor service sentiment on the quality prompt. AI holds both frames without reconciling them.

Deliveroo freshness is prompt-sensitive. Direct exit questions are accurate (March 2026). Category ranking prompts — especially Google AI Mode and Perplexity — still list Deliveroo as a live option in a large share of answers. Stale knowledge survives unless the prompt forces a status check.

SEA country context is uneven. Gemini localises better (GoFood/Gojek for Indonesia, LINE MAN for Thailand). ChatGPT often returns near-identical regional Grab answers for SG and MY prompts.

Citation landscape

Domain pattern Role in Week 1
misslobang.com / divedeals.sg / shopback.sg Deal and subscription comparison framing
reddit.com Service quality and Grab vs foodpanda sentiment
grab.com / food.grab.com Official speed and coverage claims
straitstimes.com / channelnewsasia.com Deliveroo exit and reliability news
momentum.asia SEA market-share narratives

Implication for AEO: affiliate comparison blogs and Reddit shape food-delivery AI answers as much as brand sites. Subscription economics (pandapro vs GrabUnlimited) are a recurring decision framework engines copy.

What this index is not

  • Not a coupon guide. We track recommendations, not live promo codes.
  • Not verified market-share research. Figures like Grab ~69% / foodpanda ~24% appear in AI answers (notably Gemini) and should be treated as claimed, not audited.
  • Not a Deliveroo post-mortem. Exit handling is noted because it exposes AI freshness gaps.

Run your own baseline

Request a free AI visibility audit to test whether your brand appears on the prompts that drive food, logistics, or super-app shortlists in Singapore.