SINGAPORE: Singapore firms prefer Retrieval Augmented Generation (RAG) and Small Language Models (SLMs) over generative artificial intelligence (AI) tools like chatbots, Large Language Models (LLMs), and digital assistants, according to ABBYY, an intelligent automation company.
According to Singapore Business Review, the report found that only five in 10 businesses in Singapore use chatbots, LLMs, and digital assistants compared to its global peers. This is lower than the United Kingdom, where about 70% of businesses have adopted such technologies.
However, Singapore leads on using RAG (37%) and SLMs (37%).
The report also found that Singapore leads in AI adoption, with just 1% of companies lacking a clear AI strategy.
In the past year, firms in Singapore invested an average of US$1.3 million (S$1.72 million) in AI.
About 63% of firms mentioned customer expectations as a key reason for using AI tools, while 97% of businesses said they have significant trust in the government regulations on AI tools.
Almost four in ten (37%) Singapore firms said they implement AI in specific areas of their business, with a similar percentage applying it broadly across most or all operations.
However, only 22% of firms use AI in limited functions or departments.
According to AWS, RAG optimises the output of LLMs by referencing an authoritative knowledge base before generating responses. This method improves LLM output for specific domains or an organisation’s internal knowledge without requiring retraining, making it a cost-effective way to ensure accuracy and relevance.
Meanwhile, SLMs are similar to LLMs but with less training data and parameters. “SLMs are ideal for specialised, resource-constrained applications, offering cost-effective and rapid deployment capabilities,” according to TechTarget. /TISG