AI in Malaysia · Automotive

820,752 new cars, about 870,000 used: where AI helps a Malaysian car dealer, and where it must not

Car sales are at record levels, hire purchase changed on 1 June 2026 and used car buyers are taking mileage disputes to the tribunal. For a dealer, the useful AI answers the enquiry, reads the paperwork and checks a car against its own records. It does not quote the instalment.

Pikabyte Engineering 7 min read

870,000

used vehicle transactions in Malaysia in 2025, as estimated by the used car dealers' federation FMCCAM

Malaysians bought 820,752 new vehicles in 2025, according to the Malaysian Automotive Association (MAA). It was the second year in a row above 800,000. The first half of 2026 was up another 3.1%, and in July the MAA raised its forecast for the year to 800,000 units.

The used market is about the same size. The Federation of Motor and Credit Companies Associations of Malaysia (FMCCAM) estimates about 870,000 used vehicle transactions in 2025 and expects 850,000 to 900,000 in 2026.

Volume is not the problem. What a dealer manages is everything around the sale: the enquiry at 10pm, the loan documents, the trade-in, the ownership transfer and the service visits afterwards. Several of those changed in 2026.

What changed in 2026

Hire purchase. The Hire-Purchase (Amendment) Act 2026 came into force on 1 June 2026. It abolishes flat interest rates and the Rule of 78. Interest is now calculated on a reducing balance and quoted as an effective interest rate. Agreements can be signed and delivered electronically. Bank Negara Malaysia says banks have a transition period until 31 March 2027 to upgrade their operations. Every salesperson’s habit of quoting a flat rate from memory is now out of date.

Electric cars. Malaysia recorded 58,129 electric vehicle registrations between January and October 2026, up 85.9% on a year earlier. FMCCAM said in January that used EV prices have softened because of aggressive new EV pricing and faster depreciation. A trade-in valued on last year’s assumptions is a loss waiting on the forecourt.

Ownership transfer. On 25 May 2026 the Road Transport Department (JPJ) suspended online ownership transfers after complaints that vehicles had been transferred without their owners’ knowledge. The service was restored the same day. The episode is a reminder that a transfer is only as good as the record of who agreed to it.

Buyers who complain. The Tribunal for Consumer Claims recorded 1,622 vehicle and automotive cases between 2020 and mid-2026, with awards of about RM3.9 million. In September 2026 the tribunal in Johor ordered a used car dealer to pay RM5,000 to a buyer whose car was shown with 80,000 km on the odometer. A service centre’s records put it at 370,000 km.

A lemon law is coming. Budget 2026 announced amendments to the Consumer Protection Act 1999 to introduce one. As of mid-October 2026 they had not been tabled in Parliament, and what a dealer will owe a buyer is not yet defined.

Where AI is useful today

1. Answering the enquiry first. A buyer messages three dealers about the same model. The one who replies first with the real price and a viewing time usually gets the visit. Carsome’s chief business officer told Utusan Malaysia in November 2025 that its AI call centre, Emily, handles eight in ten customer calls. A single dealership does not need a call centre. It needs an assistant on its WhatsApp number that answers from the actual stock list, books the test drive and passes the chat to a salesperson with the details collected. We set out the rules and costs in AI agents on WhatsApp.

2. Reading the loan file. A loan submission is an identity card, a driving licence, payslips, bank statements and an EPF statement, usually sent as photos. A model can read these into one record and say what is missing before the file goes to the bank. A person confirms the figures. We use the same pattern in production for bank statements of account.

3. Checking a car against its own records. This is the most valuable job on the used side, and it is done before you buy the stock, not after you sell it. The Johor buyer found the true mileage in two places: a service centre’s system, and a vehicle history platform that showed an insurance claim in August 2023 at 337,149 km. A dealer can run the same comparison on every car it takes in. The model reads the service book, the inspection report and the history report, lines up the dates and readings, and flags any that go backwards.

4. Writing the listing from the record. Once the inspection is a structured record, a model can turn it into a clear advertisement for each portal in seconds. The rule is that it writes only what is in the record. A sunroof the car does not have, in an advertisement with your name on it, is your problem.

5. Reminding owners to come back. Service due dates, road tax and insurance renewals are dates you already hold. A reminder that offers two booking slots keeps the customer with your workshop.

What it should not do

Quote an instalment. The monthly figure and the effective interest rate are arithmetic under rules that changed on 1 June. They belong in the bank’s calculator or in tested code, with the result shown to the customer as a quotation. A language model asked to work it out will give a confident number that may be wrong.

Vouch for a car. “Accident free” and “low mileage” are statements a tribunal can test. If the record does not support it, the assistant should not say it, and nor should the salesperson.

Decide who gets a loan. A model can check that a file is complete. It should not score a person. Guidelines issued by the Personal Data Protection Commissioner on 30 April 2026 say AI should not be relied upon as the sole basis for a decision about a person, and expect people to be told when automated decisions are used.

Set the trade-in price alone. Carsome uses AI to generate real-time pricing, with the scale and data of a regional platform behind it. For most dealers a report of comparable listings and days in stock is the realistic version, with the buyer making the call.

The workshop

AI does not close the skills gap in the service bay. In May 2026, Free Malaysia Today contacted ten independent workshops in the Klang Valley, and seven said they do not carry out EV repairs because they lack specialised training. In August, Pro-Net, which sells Proton’s e.MAS electric cars, recognised its first 37 certified EV technicians, drawn from 25 dealerships.

Software helps around the edges. A service adviser’s notes become a clear job card. A customer gets a plain explanation of what was done and why. The repair itself still needs a trained person.

Customer data

A dealership holds identity cards, payslips and bank statements for every loan applicant, including the ones who were rejected.

  • An AI model is a third party. Know where the provider processes documents and how long it keeps them. We covered this in the amended PDPA for software teams.
  • Delete what you no longer need. A folder of old loan files on a salesperson’s phone is the most likely breach you will have.
  • Honour opt-outs. Under the PDPA a person can require you in writing to stop using their data for direct marketing. A service reminder is one thing. A weekly promotion is another.

A sensible first project

  1. Put the stock list in one place, with price, year, mileage, inspection result and photos.
  2. Let an assistant answer enquiries from it, in draft mode first, with salespeople pressing send.
  3. Add the loan file checklist. Measure how many submissions go to the bank complete the first time.
  4. Add the records check for every used car before purchase.
  5. Move the instalment calculation into code that follows the new rules, and stop quoting from memory.

The dealers who gain from AI will be the ones whose answers are quick and whose claims about a car hold up when someone checks.

Sources

This article is general information and is not legal advice.

Written by

Pikabyte Engineering

Pikabyte builds custom business systems, applied AI and secure cloud infrastructure from Malaysia. Talk to us.

The work behind this article

Platform · AI agents · WhatsApp commerce

In development

A team of AI staff that answers, books, sells and chases payment on WhatsApp, with people in charge.

A multi-tenant platform for small and medium businesses whose customers reach them on WhatsApp. AI agents answer every chat from the business's own knowledge, book appointments, score leads and follow up on invoices, and hand over to a person when one is needed. Behind the inbox sits the system that runs the business: CRM, calendar and rosters, jobs and quotations, stock, invoicing and automation.

8
AI staff roles, from front desk to owner's assistant
589
automated backend tests
5
integrations: Google, Shopify, TikTok, Lalamove, Do Not Call registry
1
gateway every AI call passes through, metered for cost
Tour AllSet Northridge Technologies · Singapore

Start a conversation

Want this applied to your business?

Tell us what is slowing the business down. You will talk to an engineer, not a salesperson, and get a straight answer on whether software is the right fix.