Grab Seeks AI Edge in Product Development Using Agentic Engineering, Vibe Coding

Why It Matters

Accelerating development cycles and extending AI‑assisted capabilities beyond engineers lets Grab launch features faster, sharpening its competitive edge in the crowded Southeast Asian super‑app market.

Key Takeaways

• Grab's "Virtual Store Manager" prototype built via AI-driven vibe coding.
• Over 90% of Grab engineers now use AI coding tools daily.
• AI agents boost merge requests 40% per engineer, cut development time 20‑30%.
• Agentic engineering lets AI handle tasks, humans review final output.
• Grab aims to become a “cyborganisation” blending humans and AI agents.

Pulse Analysis

Artificial intelligence is reshaping how super‑apps build products, and Grab is at the forefront with its "vibe coding" approach. By prompting an AI to generate code, non‑technical staff can spin up high‑fidelity prototypes in days rather than weeks, dramatically lowering the barrier to innovation. This method mirrors design‑first workflows, where ideas are validated with real users before heavy engineering investment, a practice that aligns with agile product strategies common among leading tech firms.

Beyond rapid prototyping, Grab’s "agentic engineering" model assigns entire development tasks to AI agents, reserving human intervention for final verification. The shift has already yielded measurable gains: more than 90% of engineers now rely on AI tools daily, merge request volume has risen roughly 40% per engineer, and overall development timelines have shrunk by 20‑30%. These efficiency lifts not only expand engineering bandwidth but also enable Grab to iterate on features for its massive Southeast Asian user base—over 20 billion rides and orders provide a data moat that fuels AI‑driven insights.

However, the speed boost carries cultural risks. Grab’s CTO warns of a "normalisation of deviance" as AI‑generated code proliferates, underscoring the need for strict governance and testing standards. Maintaining disciplined code reviews ensures that the rapid output does not compromise reliability, especially at Grab’s scale. As the company pursues its "cyborganisation" vision—seamlessly integrating humans and AI agents—it sets a benchmark for how large, multi‑service platforms can harness generative AI while safeguarding quality and trust.