Artificial intelligence (AI) is a key driver for Grab, with over 1,000 AI models embedded across its platform. Guided by the principle of “AI-first with heart,” the company aims to translate AI into real value for users, partners and communities across Southeast Asia.
While many companies in the region are still testing AI agents internally, Grab is placing them directly in the hands of its merchants and drivers. These AI agents, which are software systems capable of making decisions and acting autonomously, go beyond passive analysis. Developed in collaboration with OpenAI and Anthropic, Grab’s agentic tools serve as intelligent assistants that help partners streamline operations, make smarter decisions and improve productivity.
Grab’s AI Driver Companion, for instance, uses real-time and historical data to forecast ride demand and guides drivers to high-demand zones with greater precision, eliminating the guesswork and helping them secure jobs faster. The feature is used by more than 250,000 drivers every week across Grab’s markets.
Another example is an AI-assisted voice reporting tool that lets drivers verbally report traffic conditions such as jams, roadworks or flooding. Supporting multiple languages, it replaces manual reporting and allows drivers to update Grab’s map safely while on the move.
Since its pilot rollout, the tool has logged over 16,000 reports daily from more than 900,000 drivers, enhancing the accuracy of GrabMaps and improving overall route efficiency.
Personalised business advisor
See also: SaladStop! and Razer tap on generative AI to level up nutrition and gaming
Merchants are receiving similar support. Integrated into the GrabMerchant app, the AI Merchant Assistant offers personalised business advice around the clock through a chat interface. It responds with tailored, human-like recommendations to a wide range of business and operational queries, from boosting sales to optimising menu layouts or launching advertising campaigns. It can also proactively suggest improvements based on performance trends.
This feature also helps boost Grab’s revenue. “The AI Merchant Assistant interacts with sellers as if they are humans. And we’re seeing sellers spending 24% more on advertising on our platform when they use the bot,” CFO Peter Oey says in an After Earnings podcast.
Powered by large language models (LLMs) that blend general knowledge with merchant-specific operational and transactional data on Grab, the assistant delivers highly personalised insights to aid business growth. In the coming months, it is expected to go a step further, recommending financing options, including loan products from GrabFin and its digital banks, based on a merchant’s profile and performance history.
See also: Alphabet CEO expects to keep hiring engineers as AI advances
Grab is also using AI to elevate how merchants sell. Through an AI-powered sales coach that is built on Salesforce’s Agentforce platform, sellers can receive on-demand training that adapts to their specific products, customer segments and sales opportunities.
The tool guides merchants through tailored role-play scenarios and evaluates their sales pitch in real time, referencing their customer relationship management data. Whether refining a pitch or preparing for a negotiation, the AI coach delivers precise, contextual feedback to help improve conversion and maximise return on investment.
AI for consumers and internal teams
On May 23, Grab launched an AI Centre of Excellence (COE) at its GrabHQ@one-north campus in Singapore, with a mission to accelerate AI-driven solutions that improve accessibility, boost productivity and contribute to Southeast Asia’s broader digital transformation.
One of the COE’s first innovations is a voice assistant designed for visually impaired users. Currently in pilot in Singapore, the feature enables users to book rides via voice commands and is being tested in collaboration with the Singapore Association of the Visually Handicapped.
Built on OpenAI’s models, Grab has fine-tuned the speech-to-text engine using 80,000 local voice samples and point-of-interest names, improving recognition accuracy of Singaporean accents and landmarks from 46% to 89%. A voice donation feature is set to launch in June, inviting users to contribute samples that will further train and refine the model.
The AI COE is also focused on internal enablement, giving Grab employees tools to prototype and scale AI ideas quickly. Its in-house LLM kit (containing pre-built code and advanced AI models) allows its internal teams to spin up secure testing environments in under a day, as compared to the 1.5 weeks it previously took.
To stay ahead of the latest tech trends, click here for DigitalEdge Section
Among the AI solutions for internal teams is Mystique, an in-house AI-powered copywriting tool. Trained on internal data and brand tone, Mystique has slashed content creation time by 98.5% and boosted engagement rates by up to 50% across channels such as push notifications, email campaigns and in-app messages.
Another internal application is Jarvis Report Builder, an AI-powered tool that automates the creation of merchant performance reports. Previously compiled manually by account managers over six to 10 hours each week, these monthly reports can now be generated in minutes. By entering basic details such as the merchant’s name, country of operation and peer benchmarks, Jarvis produces a customised performance summary and slide deck, enabling more focused and data-driven discussions with key merchants.
Since its launch in July 2024, the AI tool has saved Grab more than 30,000 hours in manual work, the company said in November. Grab has since developed a suite of software development kits that extend Jarvis’s capabilities to cover additional report types, including market trend analysis.
Scaling up with cloud ERP
While AI takes centre stage in Grab’s transformation, the company is also strengthening its financial foundation with Oracle Cloud Enterprise Resource Planning (ERP) as it expands across Southeast Asia.
“We rely on Oracle Cloud ERP to streamline and manage our financial operations. From core financial management and procure-to-pay automation to order-to-cash, multi-entity compliance support, real-time reporting and ensuring SOX
(Sarbanes-Oxley Act) compliance, Oracle Cloud ERP empowers us with full control and visibility across our business,” Quinny Lei, Grab’s head of Business IT Solutions, tells The Edge Singapore.
Oracle Cloud ERP supports core finance functions including general ledger, payables and receivables, fixed assets and cash management. This enables Grab to maintain real-time financial data and streamline month-end close processes across markets. It also automates procurement processes and monitors partner and merchant activities (particularly across GrabFood, GrabMart and GrabAds), helping ensure timely revenue recognition and accurate reconciliation.
In addition to these capabilities, Oracle Cloud ERP has provided Grab with a flexible, scalable system to support its rapidly expanding ecosystem. Lei adds: “The platform’s open application programming interfaces (APIs) enabled flexible integration with new products, quickly connecting new internal systems and third-party platforms as we launched verticals (like insurance, investments and B2B services) without disrupting existing operations. This allows us to tailor workflows, approval chains and reporting hierarchies for each vertical while still consolidating data into a unified financial view.”
With operations across multiple jurisdictions, Grab needed a financial system capable of managing regulatory complexity. Oracle’s support for multi-entity, multi-currency configurations, combined with localised tax and compliance features, offered the flexibility to comply with local laws while maintaining centralised control.
The system’s embedded dashboards and analytics give Grab’s finance team real-time visibility into financial health, allowing for performance tracking, trend analysis and more informed planning. SOX compliance is reinforced through audit trails, role-based access controls and workflow approvals.
“In short, Oracle Cloud ERP gave us the agility, control and insights to scale our superapp vision — from transport to food, payments and beyond — while maintaining operational discipline across diverse markets,” says Lei.
As it plans for the future, Grab is exploring ways to advance its finance function using predictive analytics, intelligent automation and AI-driven forecasting powered by Oracle Cloud Infrastructure and Oracle Fusion Cloud Enterprise Performance Management.
