Google Cloud launches PanyaThAI to drive agentic AI in Thailand

Google Cloud has introduced PanyaThAI, a programme focused on accelerating the use of agentic artificial intelligence within Thai organisations across sectors such as financial services, retail, education, manufacturing, media, and real estate. The programme has begun with 15 member organisations, including Bitazza, Chulalongkorn University, Dhipaya Group Holdings, Finnomena, Ocean Life Insurance, SE-Education, Shop Global E-Commerce Company, Siam Piwat, Sansiri, Skooldio, the Stock Exchange of Thailand, Thai Wacoal, TISCO Financial Group, TOPS, and True Digital Group.

Economic potential

Recent research from consultancy Public First suggests that more effective adoption of AI solutions could contribute up to THB 730 billion (USD $21 billion) in additional economic value to Thailand by 2030. However, the country still faces three key barriers: ensuring AI delivers factual, trustworthy outputs, a lack of AI-ready data, and insufficient local expertise in data management and AI.

Programmeme framework

PanyaThAI offers a comprehensive model that includes Google Cloud's AI infrastructure, research, and foundational models. The programme partners with consulting and implementation companies including Accenture, Deloitte, Digithun Worldwide, HoriXonT8, MFEC, NTT DATA, Skooldio, and Tridorian.

NTT DATA has committed to expanding local skills by certifying 300 additional Thai professionals in areas such as data analytics, AI, cybersecurity, and application modernisation to support the initiative.

Blueprint for adoption

The programme enables organisations to co-create AI adoption roadmaps tailored to business priorities. It emphasises responsible AI principles, clear governance, and aligning AI outputs with verified corporate data sources. The framework also includes support for identifying high-impact AI use cases, developing interoperable solutions, and defining ROI metrics throughout the development lifecycle. Upskilling is a priority, with comprehensive training offered to employees at all levels.

"Globally, more than half of surveyed executives reported six to 10% revenue growth from putting enterprise AI solutions into the hands of their teams and the people they serve. Their organisations are now channelling at least half of their AI budgets toward a unified agentic platform that will serve as the core engine for redesigning operational processes and strengthening their market leadership," said Annop Siritikul, Country Director, Thailand, Google Cloud.

"More specifically, companies that have embraced AI from Google Cloud have transcended 'pilot purgatory.' They have achieved an average return on investment (ROI) of 727% in three years, with a 'payback period' of eight months, reinforcing that not all AI technology providers and transformation approaches are equal. Through PanyaThAI, we're bringing a proven blueprint to organisations, providing them with full-stack AI services and essential training and expertise. Our ambition is to help them foster a workforce that's 'bilingual' in their domain and AI, accelerating their ability to solve complex challenges and drive consistent ROI from AI. We're seeing strong interest in this programme and look forward to welcoming and supporting more organisations beyond the initial 15," said Siritikul.

Early use cases

Several founding participants have already begun deploying agentic AI through the programme. SE-Education, a publisher and distributor, worked with Digithun Worldwide to launch an AI-powered semantic search for its digital marketplace. The solution helps users locate relevant educational content by intent rather than simple keyword matching. The tool uses models such as Gemini Embeddings and Gemini 2.5 Flash to match product listings with user queries, even when they do not directly overlap in language or keywords.

"We partnered with SE-ED to successfully transform their platform from a traditional eCommerce website into a comprehensive e-marketplace through application modernisation on Google Cloud. This architectural overhaul establishes the high-performance digital foundation required to support complex AI workflows. Building on this, we designed SE-ED's AI search agent to understand concepts, not just keywords. It starts by using Google's Gemini Embeddings model to curate a selection of products that are conceptually relevant to the user's query-for instance, 'The Lord of the Rings' novels in response to 'a hero destroying a powerful ring in a volcano,' even if the search terms aren't in product descriptions. This process is grounded in SE-ED's catalogue to ensure the agent never recommends 'made-up' products. The selection is then handed off to the Gemini 2.5 Flash model, which applies deep reasoning to rank and organise the order in which the products are presented, delivering final search results aligned with the user's true intent," said Niwat Chatawittayakul, Managing Director, Digithun Worldwide.

"We've established an omnichannel ecosystem to foster intellectual curiosity and a reading culture across the nation. As the market leader, SE-ED is committed to pioneering a new era where an intelligent AI agent serves as the ultimate personal reading assistant for every Thai. Building on this early success, we're looking to embed more helpful and high-value AI capabilities into this platform, as well as our SE-ED Reader and e-Library by SE-ED mobile apps," said Rungkan Paisitpanichtrakul, Managing Director, SE-Education.

"By harnessing the same powerful AI components that underpin Google Search, our AI search agent helps e-marketplace users discover highly relevant resources faster and with greater precision. This enhanced experience has proven that faster discovery leads directly to purchases-driving organic conversion rates from 12% to a remarkable 27%, while successfully reducing bounce rates to 10% and cart abandonment to just 6%," said Passpunnee Mahayos, Head of Digital Business, SE-Education.

AI for fashion

Thai Wacoal, an apparel manufacturer, is using generative media models from Google Cloud to address the need for new product images in the fashion eCommerce sector. The company worked with Tridorian to develop an agent capable of generating photorealistic images and videos for every colour variation of its products from a single original image, reducing the need for repeated photoshoots and accelerating online launches.

"Google's generative media models have mastered the physics of fabric and the 3D geometry of the human body, preserving lighting, folds, and texture while modifying pixel colour data. We've also grounded the models' outputs in Thai Wacoal's database of colour swatches and standards. This ensures the agent doesn't 'guess' a shade but retrieves exact manufacturing specifications, creating a verifiable 'digital twin' that matches the actual product. Our solution will empower Thai Wacoal to launch and monetise multi-colour apparel collections on its eCommerce platforms with minimal delay, translating high-speed, high-fidelity visual asset creation into revenue opportunities," said Andhika, Chief Revenue Officer, Tridorian.

"When our creative agent goes live in Q1 2026, it will do more than just solve our photoshoot predicament and accelerate time to market; it will offer the opportunity for Thai Wacoal to shift toward a high-mix, low-volume or made-to-order manufacturing model, where production can be tied directly to actual customer engagement and demand. Across our portfolio, we have already implemented other AI innovations, including Google Cloud's Virtual Try-On capabilities, to unlock new strategic advantages through digital realism and mass customisation," said Dr. Supranee Auiyasathian, Deputy Director of Digital Business for Wacoal, Thai Wacoal.