Author reels in $200M for its generative AI toolkit

Artificial intelligence startup Author Inc. today disclosed that it has closed a $200 million Series C investment at a $1.9 billion valuation. 

Premji Invest, Radical Ventures and ICONIQ Growth led the round. They were joined by greater than a half dozen other backers including the enterprise capital arms of Salesforce Inc., Adobe Inc., IBM Corp. and Workday Inc. The raise comes a few 12 months after a $100 million Series B round that reportedly valued Author at $500 million to $750 million. 

San Francisco-based Author develops a series of enormous language models geared towards the enterprise market. The corporate’s most advanced model, Palmyra X 004, debuted last month. The LLM significantly outperforms GPT-4o at tool use, meaning it more reliably performs tasks that require interacting with external applications resembling databases. 

Palmyra X 004 supports prompts with as much as 128,000 tokens. A token is a unit of knowledge that holds a number of characters or numbers. The LLM can write text, generate code and analyze user-provided datasets for useful patterns.

Author offers Palmyra X 004 alongside a variety of domain-specific models. One LLM, Palmyra Fin, is optimized for financial tasks resembling identifying investment risks and explaining market trends. One other LLM called Palmyra Med guarantees to hurry up healthcare professionals’ day-to-day work. 

Author provides its LLMs alongside an offering called Knowledge Graph. It’s a RAG, or retrieval-augmented generation, service that allows the corporate’s models to include data from external sources into their prompt responses. Knowledge Graph may give an AI access to information in PDFs, presentations, charts and other files.

Under the hood, the service uses an LLM to discover how the records provided by a user are connected to at least one one other. Author says that this approach makes AI models powered by Knowledge Graph less vulnerable to hallucinations. In keeping with the corporate, the service also makes the LLMs it powers more proficient at answering complex questions.

“Our reasoning engine powers AI apps which have structured inputs and outputs, allowing you to suit Author seamlessly into an existing workstream,” Author co-founder and Chief Executive Officer May Habib wrote in a blog post today. “It also powers agents that may work with other systems, think through complex tasks, and analyze your organization data.”

Author provides application programming interfaces that enable developers to integrate its LLMs and Knowledge Graph offering into their services. For business users, the corporate offers a no-code tool that makes it possible to create AI applications with none programming. Users can assemble their application’s interface from prebuilt constructing blocks and customize details resembling the style through which prompt responses are generated. 

To assist customers quickly put its LLMs to make use of, Author provides greater than a half dozen prepackaged AI apps. Some are general-purpose tools geared towards tasks resembling generating code. There are also more specialized applications, including one which can analyze marketing materials produced by financial institutions for regulatory compliance issues.

Author will use the proceeds from its newly announced funding round to expand its feature set. As a component of the trouble, the corporate plans to release more AI applications for the healthcare, retail and financial sectors. Author also plans to expand its LLM lineup with models which might be more adapt at reasoning tasks.

Image: Author