Executive summary

Use-case map cells by maturity tier
Malaysia online marketplaces · 9 process areas x 4 lenses = 36 cells · Oct 2026
Proven 4 search, ads, sorting, service Scaling 11 listings, trust, delivery Experimental 14 live hosts, payments No Malaysian case 7 halal and SIRIM checks, returns
Source: [Kira estimates] from the use-case map in chapter 2

The answer in three lines

The platforms took the first gain. Shopee's chatbots handle about 80% of customer queries and cut cost per contact by about 30% [Sea Q1 2026]; AI ad tools lifted its ad revenue by more than 70% [Sea Q2 2026].

Sellers lag. Malaysian online sellers use AI in 26% of their operations against 37% across the region, and 64% cite cost and time as the barrier [Lazada-Kantar 2025].

The rules are moving to the listing. Cabinet wants SIRIM and halal checks on platforms and a review of foreign platforms without a local base [Cabinet Aug 2026].

Value at stakeWe size AI at about US$0.4-1.0 bn a year by 2030, 1.2-2.8% of base-case GMV, and platforms keep over half [Kira estimates].
A Japanese or Korean company should sell trust, compliance and Malay-language tools to sellers and platforms, not build a fourth marketplace.
Cabinet Aug 2026 = NST and Malay Mail · Lazada-Kantar 2025 = Lazada and Kantar · Sea Q1 2026 = Sea Q1 2026 deck · Sea Q2 2026 = Sea Q2 2026 results · Kira estimates = KIRA analyst triangulation

Marketplaces sit in the most-exposed corner: the whole offer is digital and every step leaves data

Exposure combines how information-heavy the work is with how much of the offer can be delivered digitally; staff cost decides how fast it pays.

AI exposure by industry in Malaysia
KIRA scores 1-5 · information intensity vs share of the core offer that can be digitised · 2026
Information-rich, hard to digitise Most exposed Least exposed Digital, less data-heavy Share of the core offer that can be digitised (1-5) Information intensity (1-5) Online marketplaces Retail banking Telecoms Logistics Store-based retail Restaurants Manufacturing every order, search and refund is data
Source: [Kira estimates]

Why the score is high

The market is large and fully online. E-commerce reached about US$20 bn in 2025, up 21% [e-Conomy SEA 2025], with 98.3% of Malaysians online [DOSM ICT 2025].

Recommendations already steer buying. 68% of Southeast Asian consumers say AI-generated recommendations influence them [e-Conomy SEA 2025]; this is a regional figure, not a Malaysian one.

CaveatHigh exposure does not mean equal gain: work done by people at low wages, such as packing and live hosting, pays back slowly.
Treat AI as part of the platform's operating model, not a feature: the question is which layer of the chain captures the gain.
DOSM ICT 2025 = DOSM ICT use 2025 · e-Conomy SEA 2025 = Google, Temasek, Bain · Kira estimates = KIRA analyst triangulation