Theta Lake

Theta Lake captures collaboration tools more completely than any peer we cover. Its 100+ certified integrations capture video, screen-share, whiteboards, files and AI-assistant prompts and outputs in context, with an SEC 17a-4 archive and 85+ AI detections. Zoom and Webex sell their compliance products on Theta Lake, Gartner ranked it first in five of six use cases in its 2025 Critical Capabilities, and it added 140+ customers in 2025.

Scale and disclosure are the limits. Theta Lake has raised about $71m, last in March 2022, and has about 110 staff. ISO 27001, SSO and SCIM, a public no-training statement and pricing were not located, its SOC 2 type is gated, and no accuracy benchmarks are published. It has no trade surveillance, and its review agents keep humans in the loop.

We initiate at 6.87, Established, with an Improving outlook. Data & Content, at 8.00 on the segment variant, is the highest in the segment, and Theta Lake is the only company in the Established Solutions quadrant. Commercial Value, at 4.50, is the weakest. Pricing disclosure and identity controls would move the score most.

Strengths

• Collaboration capture. Video, screen-share, whiteboards, reactions, files, chat, voice and AI artefacts captured in context and replayable.
• AI-assistant governance. Capture and inspection of Copilot, Zoom AI Companion, Claude and ChatGPT Enterprise, with prompt-injection and jailbreak detection.
• OEM distribution. Powers Zoom Compliance Manager (2024) and Webex Compliance Hub (2026).
• Analyst validation. First in five of six Gartner Critical Capabilities use cases (2025); Visionary in the Magic Quadrant.
• AI assurance. ISO 42001 and CSA STAR for AI Level II certification; classifier audit and drift reports.
• Integrations. 150+ API endpoints, Relativity, CrowdStrike and data-warehouse connectors.

Cautions

• Scale. About $71m raised, no round since March 2022, and about 110 staff.
• Assurance gaps. ISO 27001 not located; SOC 2 type gated; SSO and SCIM not confirmed.
• Training terms. No public no-training statement; models are trained partly on customer feedback.
• Evidence. No published accuracy benchmarks; capability counts are company figures.
• Scope. No trade surveillance and no MNPI module.
• Pricing. Not published; no trial.
• Autonomy. Review assistance rather than autonomous triage.