Vellum AI

Generative AI Prompt Engineering Startup Raises $5M

Prompt engineering startup has raised $5 million in a seed funding round from a group including Y Combinator, Rebel Fund, and others. As more companies start employing generative AI tools to generate text and other content, they may struggle with composing requests for the large language model-powered AI. Vellum provides solutions to craft better prompts for those tools.

The capabilities of ChatGPT, Google Bard, and other generative AI products rely on users crafting their questions, tasks, and other input in a way that they can understand not only the basic meaning but the subtext and larger meaning behind the phrases. It’s something humans do automatically, but just because the AI understands “natural language” doesn’t mean it grasps it as well as a real person. Much like learning a computer coding language, designing a generative AI prompt requires thinking about how the software processes human demands.

Vellum Prompting

LLM developers and businesses integrating generative AI into their systems need to develop and test prompts to discover what works best and how to improve what they’ve already designed. Vellum helps streamline the road to a useful prompt and allows users to add data from their company to flesh out the context behind a request and see how different models respond in a side-by-side comparison. Since its launch at the beginning of the year, Vellum claims to have acquired more than 50 paying customers, though the exact business model is opaque.

“When we announced Vellum 5 months ago, our idea was inspired by the pains we had personally felt while developing applications on LLMs since mid 2020,” Vellum’s founders wrote in a blog about the new funding. “The Generative AI industry is evolving rapidly (new models from Anthropic, OpenAI, Inflection AI, Google, Mosaic etc.) and this is just the beginning; over time there will be widespread adoption of AI in companies around the world and we intend to be there to enable it. While our platform has evolved significantly over the last few quarters, our mission remains the same: help companies build production use cases with Large Language Models. We’re excited to use this funding to double down on this mission.”

Prompting Funding

There’s a growing place for companies that can assist other enterprises in applying generative AI and investment to match. French startup Mistral recently raised $113 million for building and training large language models to support enterprise generative AI projects, while  Inflection AI has convinced big names like Nvidia and Microsoft to bet big, $1.3 billion big, on its plans to build the largest AI cluster in the world one that can power LLMs that understand subtleties of human thought and emotion.

Enterprise interest has also encouraged plenty of partnerships and acquisitions, such as conversational AI provider LivePerson’s recent arrangement to bring Cohere’s LLMs to LivePerson’s chatbot platform, training them so LivePerson clients can synthesize generative AI chatbots for their needs. And Databricks was willing to pay $1.3 billion to acquire generative AI model startup MosaicML and fuse its technology into its massive data storage servers.


‘Personal’ Generative AI Startup Inflection AI Raises $1.3B From Microsoft, Nvidia, Bill Gates, and Eric Schmidt

Generative AI Safety Monitor Startup CalypsoAI Raises $23M

Generative AI Startup Typeface Raises $100M for Branded Synthetic Media Platform