A cloud-based platform has emerged that makes it easy to attach large language models (LLMs) with enterprise applications.

Using this platform, it is feasible to construct an agent that connects data sources, tools, and application programming interfaces (APIs) owned by each company with LLMs, including ‘ChatGPT’. In other words, it’s a link that helps to utilize external super-scale artificial intelligence (AI) for its own business.

TechCrunch reported on the thirtieth (local time) that US AI startup Fixie has launched ‘Fixie’, a cloud platform that permits LLMs much like ChatGPT to be connected to corporate data, systems and workflows.

Pixy principally supports ‘GPT-4’. Firms can provide their very own models or use other industrial and open source models.

The corporate initially released it as a developer preview, allowing anyone to integrate LLM into their applications, data or tools. Personal use will likely be released freed from charge.

Fix CEO Matt Welsh introduced it as “the primary enterprise Platform-as-a-Service for constructing an LLM automation platform.”

The agents it builds will be used to automate business processes, construct natural language understanding of products, answer questions on API connections and data sources, and more.

For instance, corporations can use the platform to integrate LLM functionality into customer support workflows by constructing agents that take customer tickets as input, mechanically look up customer purchases, issue refunds if obligatory, and generate responses to tickets. can do.

“Pixie permits you to wrap an existing data source or API with a little bit of code that connects to a big language model of your alternative,” said CEO Walsh Pixie. It might be converted right into a call to an external system.”

Meanwhile, the corporate announced on the thirtieth that it could raise $17 million (about 22 billion won) in seed round funding and use it to construct a team and expand the scope of the platform.

Chan Park, cpark@aitimes.com