Sign up for our daily and weekly newsletters to get the latest updates and exclusive content on top-tier AI coverage.
In the hurry to integrate AI into a company’s knowledge base, critical information can sometimes be overlooked. While it’s straightforward to upload images, videos, documents, and spreadsheets to train models, handling other unstructured data types like invoices, emails, and PDFs is more challenging. Though these files can be converted into searchable formats, the process is often manual, time-consuming, and requires machine-learning skills that many engineering teams lack.
One company, however, believes it has found a solution. Founded by Antonio Bustamante and Upal Saha, Bem is an AI data interface company that offers an API for developers to convert any input—structured or unstructured—into a desired data shape. Bem recently raised $3.7 million in seed funding, led by Uncork Capital, with contributions from Lookout’s founder Kevin Mahaffey, Roar Ventures, and angel investors like Y Combinator’s Garry Tan.
Bustamante explains that engineering teams spend about 44% of their time building, monitoring, and maintaining data pipelines. Bem aims to redefine interoperability and integration, allowing engineers to focus on essential tasks instead of worrying about data pipelines and integrations.
Bem’s service targets engineers and is designed to be user-friendly, requiring no special training or configuration. Developers can use Bem’s API to specify the data shape or schema they need before sending their information. Bustamante likens it to how Stripe started with a simple API that’s easy to implement.
Bem functions as a continuous data pipeline, converting messy data into structured formats that align with a company’s internal systems. This is what Bustamante refers to as providing Structured Data as a Service (SDaaS). Bem focuses on transforming unstructured inputs into structured data, which can be integrated directly into end products.
Bem’s potential is significant, addressing a common “hair on fire” problem where companies lack the resources or expertise to handle data transitions on their own. The company is particularly useful for industries like logistics, supply chain, healthcare, and insurance, where data needs to be structured efficiently for product development rather than document review workflows.
Built on a combination of foundational and open-source models, Bem’s system gets smarter with use, yet it doesn’t train global models on customer data, ensuring data isolation and security. Currently available through a private beta, Bem has ten early customers and is primarily targeting organizations from Series C to public stages, especially those in logistics, supply chain, healthcare, and insurance.
The seed funding will further enhance Bem’s platform, allowing investment in engineering, research, development, and product improvement. Andy McLoughlin of Uncork Capital notes that Bem aims to make data ingestion seamless, a common challenge across industries.
Bem is essentially democratizing data processing, providing tools even smaller companies can use to manage data as effectively as tech giants. While Bem has some competitors like Unstructured, which focuses only on documents, its real competition lies in the status quo of in-house data solutions. Many companies realize that building such systems internally is often more costly and complex than anticipated.
For now, Bem customizes its pricing based on each client’s data volume but plans to establish a more standardized pricing model soon. The goal is to deliver a considerable return on investment for their clients by addressing a significantly costly problem.
No valuation for Bem has been disclosed.