Israeli Startup Flexor Named "CDO Magazine Product to Watch 2024" for Pioneering Text Data Transformation

In an era where unstructured data has become the new frontier for AI innovation, a new player in the unstructured data preparation space is making waves. 

Recently named a “Product to Watch” by CDO Magazine for 2024, the platform, developed by Flexor, promises to unlock value from enterprises' vast stores of textual data. Industry experts are closely watching this technology's potential impact on how businesses leverage their textual assets.

Earlier this year, Flexor was featured on the Bessemer Venture Partners Market Map for AI Infrastructure in the 'Pre-Processing and Pipelines' category, further validating its innovative approach. This recognition highlights the company's significant role in advancing unstructured data handling within the AI ecosystem.

The “Unstructured Data” challenge

For years, unstructured data was considered pure waste — cheapest storage, never utilized, not even considered as data. However, the rise of GenAI and LLMs has dramatically shifted this paradigm, transforming unstructured data into the lifeblood of modern AI initiatives. Textual data is everywhere: Chatbots, video calls, support tickets, internal communication, emails, and more.

The 'modern data stack,' built for structured data, is facing unprecedented strain. The proliferation of unstructured data generated and the need to harness it with LLMs are pushing these systems to their limits.

Industry analysts highlight several challenges faced by data teams:

• Limitations in unstructured data processing: Capabilities are limited by outdated and low-quality keyword-based tools, or challenged by the complexity of building reliable and scalable LLM pipelines in-house.

• Getting the right data to the right downstream application: The rapid growth in the number of GenAI applications challenges existing data architectures, leading to convoluted and inconsistent data pipelines.

• Challenges in unstructured data quality: The wide variety of new data sources requires standardization and normalization to facilitate all kinds of BI, ML, and GenAI workflows.

Pioneering unstructured data management

The company was founded by Or Zabludowski (CEO) and Regev Ben Simon (CTO), both veterans of elite Israeli military intelligence units. The founders' combined experience in leveraging data to tackle complex and high-stakes challenges became the cornerstone of Flexor's mission.

Zabludowski's expertise in data analysis took center stage when he led Israel's national COVID-19 HQ Data Team. This critical role put him at the forefront of Israel's tremendous success in leveraging data to combat the pandemic.

For Zabludowski, the “aha” moment happened while trying to analyze patient data during COVID, and he realized that most of the business processes today are based on unstructured text.

Industry experts estimate that 90% of the world’s data is unstructured, spanning from sales and customer service to legal and procurement.

"We quickly realized that to process unstructured data you need a new data transformation paradigm – can you turn text into tables? Into data? So we went on a journey to bring this data to the hands of every data practitioner in the organization. Flexor stands out as the world's first data platform built for unstructured data, that integrates into existing data ecosystems," says Zabludowski.

The mission: To democratize unstructured textual data

Flexor's solution handles data throughout its lifecycle — extraction, preparation, structuring, embedding, and model management. It addresses critical challenges faced by enterprise data and AI departments, including:

• Unification, organization, and structuring of diverse textual data at scale.

• Enabling analytics and ML applications without specialized infrastructure.

• Managing complex GPU and LLMOps infrastructure for production-grade applications.

• Ensuring compliance in processing sensitive information.

• Optimizing models for company-specific terminology and knowledge.

• Modeling data effectively for both conversational data and documents.

• Seamless integration with existing data ecosystems, including data warehouses and data lakes.

By seamlessly integrating diverse data formats, this innovative platform has the potential to revolutionize how organizations harness their unstructured assets. The technology's ability to process complex data structures is being viewed as a potential game-changer in today's data-driven business landscape.

The Flexor team envisions a future where “every data practitioner can effortlessly leverage unstructured data using familiar tools, with the simplicity of SQL, without needing specialized skills or complex infrastructure.”

"We unlock a new kind of business-critical data for AI. Tribal knowledge, manual work, customer communications… are all unstructured. We are barely scraping the surface with these co-pilots. Shaping the future of data management starts with structuring the unstructured," elaborates Zabludowski.

Revealing a frequently asked question from CDOs and CAIOs, CTO Ben Simon says that it is about ways to increase the business impact of GenAI. “The answer is simple — begin by removing the data preparation barrier, then democratize access to business analysts. High-impact use cases will quickly emerge."

The CDO of one of the largest retailers in the U.S. shares that "Flexor has revolutionized our approach to handling refunds and customer satisfaction. By simplifying our data preparation process and bringing LLMs into the game, we've gained the ability to accurately analyze our refund patterns at scale and automate our playbooks. This has not only saved costs but also improved our customer satisfaction. It's transformed how we balance customer-friendliness with business sustainability."

Future outlook

The platform’s ambitious roadmap aims to make it the central hub of unstructured data processing in enterprises.

As AI continues to advance, Flexor's technology is poised to play a critical role in unlocking new possibilities for data-driven innovation by unifying all kinds of natural language sources, integrating unstructured data into decision-making processes, enabling more integration with sophisticated AI applications (like RAG and agentic systems), and creating feedback loops for AI systems. 

Flexor’s Data Residency Version is in development, which would be fully deployable within a customer's Private Cloud Environment and leverage self-hosted models. This move is seen as a response to the evolving needs of enterprise-level data management, particularly in terms of data security.

As the unstructured data management landscape continues to evolve, CDO Magazine will continue to monitor how this new platform impacts enterprise data strategies and ROI across organizations. With its innovative approach to democratizing unstructured data access, Flexor is well-positioned to shape the future of how businesses leverage their textual assets in the AI-driven world.