Staff can't code? No prob. Singapore superapp's LLM whips up apps for them

NP-hard to NP at all Southeast Asia's Uber-esque superapp, Grab, has developed a tool that allows its employees to build large language model (LLM) apps without coding....

featured-image

Southeast Asia's Uber-esque superapp, Grab, has developed a tool that allows its employees to build large language model (LLM) apps without coding. The company this week described the tool, which it calls Spellvault, as empowering "everyone in the company, even those without coding skills," to "create LLM apps that have access to the rich repository of Grab data." The superapp claimed custom apps can be built in as little as five minutes.

It gave two examples of groups that had developed tools. Its insurance division built a chatbot to provide information on policies, inquiries, and claims procedures, and its localization group developed a translation app. Spellvault operates on the front end through a three-step process.



Users start by defining the model's task, such as assessing product initiatives for potential policy risks. Then they upload relevant data into the Knowledge Vault, a repository where Spellvault stores structured information. Finally, users can specify APIs if the model needs additional data from external sources.

The platform is also used as a sharing tool among Grabbers where colleagues access existing code. On the back end, Spellvault uses retrieval-augmented generation (RAG), according to a post on Grab's tech blog last month. The AI framework integrates a retrieval step before generating responses, which is claimed to improve accuracy and relevance of LLMs.

RAG enables Spellvault to call relevant plugins like Grab's in-house Data-Arks API to answer user questions. The Python-based Data-Arks acts as middleware serving up the needed data in APIs to the LLM agents. SpellVault can pull that data from Data-Arks, which sends back SQL output, which is then summarized into other tools like Slack.

"LLMs like ChatGPT are typically trained on a broad range of publicly available information that may not be tailored to Grab's needs. By integrating the general capabilities of these models with Grab's internal data, we can develop highly customized tools that can perform Grab-specific tasks," said the company. Grab also plans to "utilize Data-Arks to provide images to GPT as GPT-4o is a multimodal model that has vision capabilities.

" Back in May, ChatGPT maker OpenAI and Grab said they'd made a deal to build and deploy AI for its users, partners, and employees. This included building customer support chatbots like the one designed by Grab's insurance business. ®.