Problem

Redactive AI addresses the data-security barrier to enterprise AI adoption. Organizations have accumulated “data access debt”: sprawling, often misconfigured permissions across unstructured documents and applications that security teams cannot realistically review and repair at scale. Traditional, application-level permissions do not reliably capture whether a person should access a particular document, sentence, image, or data chunk, while AI systems can surface that improperly accessible information at machine speed. The resulting pain is both security risk from internal data leaks and lost productivity when AI initiatives are delayed; remediation that takes months manually can be reduced to minutes.

The clearest use case is a regulated enterprise deploying an internal AI assistant. For PEXA, Redactive helped launch a permissions-aware generative-AI assistant to more than 1,000 employees in eight weeks, with the goal of improving productivity while ensuring that each employee could retrieve only the live enterprise data they were authorized to see.

Product / Service

Redactive is an enterprise AI-security platform delivered as an integrated layer around a company’s knowledge bases and AI tools. Its Permissions Assurance product uses AI-native semantic analysis to understand the meaning and sensitivity of unstructured data, identify anomalous document-level access across users and groups, automatically correct misconfigured permissions, and route ambiguous cases to the appropriate data owner. A browser plug-in also detects and blocks employees from sending company data to unauthorized “shadow AI” applications such as personal ChatGPT accounts, while providing visibility into how staff are using AI.

For AI applications, Redactive acts as a pre-retrieval guardrail between the knowledge base and tools such as Copilot, Glean, or a custom LLM. It leverages existing SSO, LDAP, and DSPM rules and classifications but adds contextual, fine-grained checks before data reaches an LLM. In the PEXA deployment, its permissions-aware retrieval engine performed real-time authorization checks and passed information to the model in small, secure segments, enabling live, permissioned retrieval without requiring the enterprise to abandon its existing security stack.

Market

Redactive operates in the emerging AI data-security and enterprise AI-enablement market, overlapping with data-security posture management, permissions governance, secure retrieval-augmented generation, and shadow-AI prevention. Its positioning is complementary rather than a wholesale replacement for LDAP or DSPM: Redactive adds semantic, document-level controls to the broader data-security stack. The nearest alternatives are therefore incumbent DSPM and data-security suites, including Microsoft Purview DSPM, as well as other AI-security vendors; the evidence available here does not establish a single named, like-for-like direct competitor.

Redactive demonstrated meaningful early enterprise traction rather than being merely pre-revenue: it identified HESTA and PEXA as flagship clients, reported two large financial-services customers scaling generative-AI use cases to more than 500 employees each, and had its platform deployed in production across RecordPoint’s financial-services customers in Australia and the United States. Funding coverage is inconsistent, with one report citing a $11.5 million seed round and another citing $7.5 million, but both identify Blackbird and Felicis among the lead investors, alongside Atlassian Ventures and Zapier. In October 2025, RecordPoint acquired Redactive in a deal whose terms were not disclosed, making the company’s current trajectory an acquisition-led continuation of the platform rather than an independent pre-revenue startup story.