Paris-based startup H, founded by former Google employees, shocked the industry last summer by raising $220 million in seed funding without releasing any products. However, shortly after the funding, the company faced concerns about its future as all three founders left due to "operational and business differences."

image.png

Despite these challenges, Company H has announced the launch of its first product, Runner H, an "intelligent agent" AI designed for businesses and developers to handle tasks such as quality assurance and process automation. Runner H is built on the company's proprietary "compact" large language model (LLM), which has only 2 billion parameters.

Company H has set up a waiting list for Runner H on its official website. CEO Charles Kantor stated that in the coming days, users on the list will be given access to an API to utilize H's pre-built agents, allowing developers to create their own agents. Additionally, using the API will provide access to H-Studio, which helps users test and manage the operation of these services.

Currently, using these APIs is free, but a paid model will be introduced later. Although the compact LLM is employed, the costs of building and running AI remain high, especially in a competitive environment. TechCrunch confirmed that Company H is raising Series A funding to support what Kantor calls "second-generation AI," contrasting it with the "first-generation AI" represented by companies like OpenAI.

Kantor mentioned that Company H has partnered with several clients in e-commerce, banking, insurance, and outsourcing to refine its product. Runner H will primarily focus on three specific application scenarios: Robotic Process Automation (RPA), quality assurance, and Business Process Outsourcing (BPO).

RPA is a field that has existed for many years, aimed at automating repetitive human tasks through basic scripts. Runner H aims to perform RPA on modified forms, websites, and templates without the need to rewrite scripts. In terms of quality assurance, Runner H can effectively reduce the maintenance burden of website testing, validating page usability, and simulating real user interactions.

BPO encompasses the enhancement and improvement of billing processes, as well as accelerating the agent's ability to access and retrieve different data sources. While the number of parameters has become a competitive focus among foundational AI companies, Runner H adopts a different strategy with its 2 billion parameters, emphasizing cost-effectiveness and operational efficiency.

Company H claims that its compact model outperformed Anthropic's "compute usage" model by 29% in the WebVoyager benchmark tests and performed well compared to models from Mistral and Meta.

Key Points:  

🔹 Company H launches its first product Runner H, focusing on "agent" applications.  

🔹 Runner H utilizes a compact LLM with 2 billion parameters, aiming to reduce costs and improve efficiency.  

🔹 The product is primarily applied in robotic process automation, quality assurance, and business process outsourcing.