Unfold unlocks legacy enterprise data for AI without migrations

Why enterprise data remains locked away

Ran Cherniavsky’s startup Unfold is targeting a critical bottleneck for enterprises adopting AI: data trapped in legacy systems, proprietary software, or applications without modern interfaces. Traditional integration tools, built for system-to-system connections rather than AI, often fail to capture the full scope of data or its business context. This leaves enterprises unable to fully leverage AI, even as board-level pressure mounts.

The Next Web reports that this gap creates immediate costs. Companies rely on specialized consultants to manually map data fields, analysts to stitch together limited reports, and AI pilots that stall because models lack access to complete datasets. Security risks also escalate when data visibility is limited, as teams cannot verify the integrity of systems they cannot read.

How Unfold unlocks and normalizes data

Unfold’s solution bypasses the need for disruptive migrations or system replacements. It works through existing interfaces—UI, reports, exports, or command lines—to extract and contextualize data within days. The company’s AI-driven approach maps data locations, relationships, and business rules, replicating the understanding of a trained employee. This includes normalizing disparate data formats (e.g., SAP’s "KUNNR" versus Salesforce’s "Account") into a single enterprise schema.

Governance is embedded in the process. Unfold tracks data lineage and ensures deterministic outputs, meaning the same query returns consistent results. This allows AI agents to act on the data without introducing variability or risk.

AI agents gain real-time access to enterprise data

Unfold complements platforms like Splunk, Snowflake, and Databricks by feeding them data from systems previously inaccessible to these tools. It also integrates directly with AI agents via connectors to Claude, Gemini, and others, enabling real-time access to enterprise data. For example, a global retailer uses Unfold to pull point-of-sale data from AS/400 mainframes into a Gemini Enterprise fraud-detection agent.

Cherniavsky’s team frames Unfold as a knowledge foundation for enterprise AI, aiming to support every AI use case within an organization. The measurable impact for customers includes time saved, cost reductions, and outcomes previously impossible, such as cross-system automation for workflows like order-to-cash.

Market growth driven by data accessibility needs

The challenge of enterprise data accessibility is urgent. Gartner projects the market for AI context platforms—tools that provide both data access and business context—will grow from $28 billion in 2026 to $78 billion by 2030. This growth is driven by the need to overcome barriers like legacy infrastructure, which still powers 45% of core systems.

Key takeaways

• Enterprises can now integrate AI with legacy systems without disruptive migrations or costly consultant-driven projects, reducing time-to-value from months to days.
• Founders building AI solutions for enterprises must address data accessibility and contextualization to avoid stalled pilots and limited model performance.
• Unfold’s approach highlights a growing market for AI context platforms, projected to reach $78 billion by 2030, signaling opportunities for startups in data normalization and governance.
• Operators in sectors reliant on legacy systems (e.g., retail, finance) can leverage tools like Unfold to enable real-time AI-driven workflows, such as fraud detection or cross-system automation.

Related on Lazyfounder

• Snowflake
• Claude
• Gemini
• SAP
• Salesforce
• More ai news
• Adobe’s AI marketing assistant now works inside ChatGPT and Copilot