Natural language processing (NLP) has become an essential part of our world’s digital transformation, and NLP startups can significantly impact the business ecosystem in Southeast Asia. NLP is part of artificial intelligence (AI) and involves extracting and analysing text and natural human language to attempt to respond to user requests and queries. An excellent example of NLP use is in search engines, whereby customer questions are analysed and put through datasets to generate a possible solution.
The rise of NLP platforms is evident, with Statista Research predicting that the market will be 14 times larger in 2025 than it was eight years ago, and revenues will reach USD 43 billion. NLP benefits businesses by boosting marketing, providing actionable insights, problem-solving, classifying texts, and generating voice and text. It also helps analyse sentiment, voice recognition, translation, enhancing customer experience, competitive edge over rivals, and more.
Can startups use ChatGPT to grow business in Southeast Asia?
ChatGPT (Chat Generative Pre-trained Transformer) is the most famous of the NLP platforms currently on the market. According to Statista Research, its global weekly interest in Google Search went up between November 2022 and 2023, hitting a peak of 100 index points. It also set a new record in the tech world by reaching one million users in less than five days.
Using generative AI, ChatGPT taps into its datasets to help startups and small and medium-sized enterprises (SMEs) to create content, respond to customer queries as a chatbot, guide customers to valuable topics, and automate repetitive tasks. Moreover, it will enable businesses to analyse data to generate summaries and save employees time to focus on non-repetitive tasks.
Top NLP platforms in addition to ChatGPT
While ChatGPT is popular, tech companies are developing alternative NLP platforms to rival it and provide more features. They include:
Google Bard
Creative and helpful collaborator, Google Bard, is a conversational chatbot designed to boost productivity and assist in generating ideas. It uses a language model known as LaMDA (Language Model for Dialogue Applications), enabling it to employ a fluid, conversational approach. Currently, Bard is experimental, meaning its responses may only partially be accurate.
Google Bard does not replicate content from other sources. As such, it will give users original content or cite a webpage if it borrows quotes from that source. Anyone with a Google account can use Bard, but it only supports US English. None of the data collected—location, IP address etc.—will be shared with third parties.
Amazon CodeWhisperer
AI coding companion Amazon CodeWhisperer enables users to build applications faster. The platform learns billions of lines of code, which helps it to generate code for applications that users are creating. It can flag unattributed lines of code from other open-source coders/programmers and scan codes to identify vulnerabilities. Furthermore, it has a selection of 15 programming languages, such as Python and Java.
According to Amazon’s research, users could complete their coding tasks 57% faster on average, and 27% were more productive.
Bing AI chat
Microsoft’s Bing has an AI chat that answers short, long, and complex questions. It gives complete answers; The more precise the question, the more detailed the response. It will summarise the response rather than give multiple confusing options. You can also prompt Bing to write poems, stories, and other text or generate images based on the keywords input.
Bing AI focuses on reliable sources and strives to be a responsible AI. That specific design means it will attempt to provide factual information and avoid sharing offensive content.
LaMDA
Google’s LaMDA (Language Model for Dialogue Applications) AI is experimental and available for review on Google’s AI Test Kitchen app. Google CEO Sundar Pichai invited users to test LaMDA and provide feedback while also understanding that the AI is still in development and may make mistakes. The company announced it had released updates to LaMDA AI to enhance safety and accuracy, but it still required fine-tuning.
Despite platforms like ChatGPT and those listed above having significant advantages, they also come with various limitations that startups should be wary of. The code developers hold substantial power to shape the AI’s actions, and there may be an element of bias in its responses.
Furthermore, the keywords used, the level of detail in the questions or prompts, and the processing capabilities of the AI may limit the results. The possibility of result errors is high, and a tech expert should ensure accuracy.
Cybersecurity and data privacy remain challenging, especially with the sheer volume of data the AI collects and analyses.
Nevertheless, NLP startups provide solutions businesses can implement to cut costs, identify gaps in company policies, get insights, and streamline operations. The platforms still require a lot of research and development (R&D) and should be considered helpful tools rather than a replacement for a real person completing the task.