Grab, a leading food delivery and rideshare company in Southeast Asia, uses OpenAI’s GPT-4 with vision fine-tuning to process millions of street-level images collected from drivers, enabling it to accurately identify traffic signs and lane dividers and reduce manual mapping labor. This approach yields more reliable, hyperlocal map data that supports their ride-hailing, delivery services, and enterprise clients.

Untukmu.AI, an online gifting site in Indonesia, uses Meta's Llama 3.1 8B model with split inference processing to protect customer privacy. By running part of the AI model on customers' devices and the rest on their servers, they deliver personalized gift recommendations without accessing or storing personal data. This ensures customer privacy while still providing high-quality, tailored suggestions, enhancing trust and satisfaction.

WRTN uses Claude to power natural, engaging character interactions and creative storytelling for millions of users across Asia. They aggregate AI services across companionship, entertainment, and productivity to make advanced AI technology accessible to everyday users in South Korea and Japan.

Grab, a leading food delivery and rideshare company in Southeast Asia, uses OpenAI's GPT-4o with vision fine-tuning to enhance their mapping data for GrabMaps. By fine-tuning GPT-4o with only 100 examples, they taught it to correctly localize traffic signs and count lane dividers from street-level imagery collected by their drivers.

Arivihan, an online learning platform in India, uses Meta’s Llama to generate personalized answers, as well as lectures, for students.

Square Enix, a Japanese video game publisher, is using Google’s AI solutions and its own customer data to develop AI-optimized marketing assets to keep its gamers engaged, sharing personalized emails suited to each player’s preferences.

Puma uses Imagen (Google’s image generator) to customize product photos on its website, ensuring they’re locally relevant across markets.

Gamuda Berhad, a Malaysian infrastructure and property management company, has integrated a Gemini-powered conversational agent into its cloud-based Tunnel Insight platform.

Gojek (a “super app” similar to Uber but for Indonesia) uses AI to allow customers to use voice commands to complete things like bill payments and money transfers.

Rakuten uses OpenAI’s models to leverage its vast data assets for improving customer service, extracting value from unstructured data, and enhancing B2B consulting. By implementing RAG (retrieval-augmented generation), they drastically improve response times for customer service, develop shopping review summarization features, and empower merchants with actionable insights.

Gupshup.io's CX Product Suite 'Converse' is now powered by ACE LLM, domain-specific models fine-tuned for enterprise-grade conversations based on Meta's Llama 2. This enhances their offerings in customer acquisition, marketing, commerce, and support.

10BedICU uses OpenAI’s API to enhance India’s critical care infrastructure. It developed tools like CARE Scribe, which transcribes doctor-patient interactions into EMR entries in multiple languages, reducing data entry time by over 50%, CARE Device Connect, which integrates data from older hospital monitors, and CARE Discharge Summary, which automates patient record summaries, improving healthcare delivery in underserved communities.

Sarvam AI specializes in developing AI solutions for Indic languages. They developed a Hindi Large Language Model by extending Meta's Llama 2 in partnership with AI4bharat. The model aims to deliver GPT-3.5-like performance on Indic languages under compute and data constraints, achieving high performance on a frugal budget.

Strategy teams at Paytm draft market landscape insights to inform their roadmaps.

Woven – Toyota's investment in the future of mobility — is partnering with Google to leverage vast amounts of data and AI to enable autonomous driving, supported by thousands of ML workloads on Google Cloud’s AI Hypercomputer. This has resulted in 50% total-cost-of-ownership savings to support automated driving.

Kakao Brain, part of Korean technology company Kakao Group, has built a large-scale AI language model that is the largest Korean language-specific LLM in the market, with 66 billion parameters. They’ve also developed a text-to-image generator called Karlo.

Tokopedia, an Indonesian ecommerce leader, is using Vertex AI to improve data quality, increasing unique products being sold by 5%.

Samsung is deploying Gemini Pro and Imagen 2 to their Galaxy S24 smartphones so users can take advantage of features like text summarization, organization, and magical image editing.

Melooha developed a hyper-personalized Vedic astrological engine using Meta's Llama 3 70B model. The model enhances the accuracy and depth of astrological insights by blending ancient wisdom with modern AI, allowing for profound exploration of astrological complexities and delivering credible, personalized predictions.

Healthify, India’s largest health platform, uses OpenAI's GPT-4 Vision and Whisper to improve food tracking and health coaching. The integration has increased food tracking by 50% and doubled user engagement with nutrition coaching. Healthify plans to expand to 20 countries and save millions of lives by improving fitness and reversing metabolic diseases globally.

Elyza, a Japanese AI startup, uses Meta Llama 2 to develop a Japanese large language model designed for advanced natural language processing applications, including language translation, chatbots, and text generation, thereby enhancing AI-driven communication and language services in Japan.

Air India integrates Microsoft Copilot for Microsoft 365, including a custom plugin for Microsoft Teams, to streamline operational data analysis and decision-making. This allows teams to quickly analyze flight performance, delays, and other key metrics through natural language queries. The solution enhances collaboration across departments, boosting efficiency, reducing operational costs, and improving the employee experience.

KissanAI has released Dhenu Llama 3, a multilingual AI agriculture assistant tailored for Indian farmers. Built on Meta’s Llama 3 8B architecture, Dhenu understands voice and text inputs and is integrated with platforms like WhatsApp, supporting 22 languages including 9 Indian languages. This assists farmers in accessing agricultural advice and information, enhancing farming practices through advanced AI capabilities.

Taiwan LLM is a pioneering large language model focused on Traditional Chinese used in Taiwan. Built on Meta Llama 2, it incorporates cultural context and advanced pre-training on comprehensive datasets, enabling it to understand linguistic nuances and cultural references, thus revolutionizing Traditional Chinese language processing in NLP applications.

Odia Generative AI is an open-source project aimed at expanding LLM capabilities for the Odia language, spoken by 40 million in India. Using Meta Llama 2, OdiaGenAI has developed the Odia Llama, a fine-tuned LLM for Odia, and is exploring applications like AI chatbots and AI tutors, thereby enhancing accessibility and digital inclusion for Odia speakers.

Digital Green uses OpenAI's GPT-4 to develop Farmer.Chat, a generative AI chatbot that supports agricultural extension services in countries like India and Kenya. It provides farmers with tailored advice by integrating Digital Green's vast library of agricultural knowledge, including training videos and research factsheets. By enhancing extension agents' ability to assist farmers, Farmer.Chat reduces the cost of traditional services and improves access to localized farming expertise.

Pratham developed 'BaalSakhi', a chatbot powered by Meta's Llama 2 (70B), integrating via WhatsApp to provide personalized, on-demand support for early childhood care and education to parents and caregivers in resource-poor settings. BaalSakhi draws insights from trusted sources like government departments, UNICEF, WHO, and childcare experts, helping users navigate their child's early development journey independently.