Zuva — AI Contract Intelligence for In-House Legal Teams

6 min read

On This Page

Zuva is a contract intelligence platform spun out from Kira Systems in 2021, offering both developer APIs and a direct contract review application for corporate legal departments.

1,300+Pre-trained AI fields
36+Business agreement types
60+Supported document formats
$20M CADSeries A funding (2021)

Overview

Zuva emerged from Kira Systems after Litera acquired Kira in August 2021, retaining 34 employees, the machine learning research team, and rights to over a decade of contract AI development. Co-founders Noah Waisberg and Alexander Hudek took the technology into an independent company, closing a $20 million CAD Series A led by Insight Partners in September 2021 with Peter Sobiloff joining the board.

The company operates under a structural constraint that shapes its entire go-to-market: under the terms of the Kira-Litera deal, Zuva serves only corporate legal departments, with law firms contractually excluded. That restriction sharpens focus but caps the addressable market. Within it, Zuva serves two distinct channels: developers and technical teams through DocAI API licensing, and in-house legal teams directly through Zuva Analyze, its generative AI-powered contract review product launched in September 2024.

The free tier Zuva launched in 2024 marks a deliberate shift in accessibility. As Waisberg explained at launch: "Contracts AI is typically pretty expensive. We've had people pay us literally millions of dollars to use a different version of the same tech. This is free." That pricing history traces directly to Kira's enterprise customer base, which at acquisition included 18 of the top 25 global mergers and acquisitions firms.

How Zuva processes documents

Zuva's processing pipeline begins with OCR using the Kofax Omnipage engine with machine learning post-processing, converting 60+ document and image formats into plain text, images, or Protobuf layout files. From there, 1,300+ pre-trained AI fields handle document classification and clause extraction across 36+ business agreement types, a model library inherited from Kira's decade of legal ML development.

The platform exposes this capability through RESTful APIs and a Python SDK, with connectors for Microsoft Copilot Studio, Logic Apps, Power Apps, and Power Automate. The Microsoft Learn connector documentation confirms support for OCR, document classification, multi-level classification, language detection, field extraction (maximum 100 fields per request), and date normalization, with a throttling limit of 100 API calls per connection per 60 seconds. Regional availability excludes US Government GCC, GCC High, China Cloud, and DoD environments.

Deployment runs either cloud-hosted or on-premise via Kubernetes. Documents are retained for 7 to 14 days before automatic expiration. Continuous learning algorithms refine model accuracy over time, and real-time collaboration supports simultaneous multi-user access with version control.

For end users, Zuva Analyze surfaces this pipeline as a direct contract review application combining generative AI and machine learning for large-scale document review. As Waisberg noted at the Analyze launch: "AI has really advanced since we sold Kira in 2021. We thought this change in technology presented an excellent opportunity to solve problems we couldn't back in the day." Beta users reported reviewing contracts 2 to 3 times faster than manual processes, though this figure is self-reported and carries no independent verification.

1

Submit documents

Upload via API, Python SDK, Power Automate connector, or browser. Supports 60+ formats including PDF, DOC, DOCX, and TXT. Free tier accepts PDFs under 150 pages or 5 MB.

2

OCR and classification

Kofax Omnipage engine with ML post-processing converts documents to structured text. Multi-level classification identifies agreement type across 36+ business agreement categories.

3

Field extraction

1,300+ pre-trained AI fields extract clause-level data. Maximum 100 fields per API request. Date normalization applied automatically. Custom model creation available without data science expertise.

4

Review and export

Analyze surfaces results in a collaborative review interface with version control. API responses return structured JSON. Free tier restricts export and copy/paste; paid tiers provide full output access.

Use cases

Zuva Analyze targets in-house legal departments handling their own deal review without routing work to outside counsel. The Artificial Lawyer product walkthrough confirmed project-based pricing to accommodate fluctuating demand, alongside a free trial accessible without a sales call. The free browser-based tier, requiring no registration, supports lease terms, vendor and supplier contracts, M&A diligence, credit agreements, employment agreements, and finance, operations, and privacy terms, with no export or copy/paste functionality.

Pricing offers both subscription and per-project options. The per-project model is uncommon among contract review competitors and targets legal teams with variable rather than continuous workloads. Teams evaluating legal document automation workflows will find Zuva's corporate-only focus a meaningful differentiator from platforms that serve both law firms and in-house departments. Recital, which also specializes in contract lifecycle management for legal teams, offers a comparable point of comparison for in-house departments weighing dedicated contract intelligence tools.

Developer integration

Before Zuva Analyze, Zuva operated exclusively as an API business, licensing its ML models to software companies and technical consultants building document processing applications. Waisberg positioned this explicitly at launch: "Customers will be other software companies or technical consultants that are building applications." Three customers signed before DocAI reached production, including Vigilant AI, with contracts at five-to-six-figure USD values.

Microsoft's use of the DocAI connector for internal legal approval workflows illustrates the enterprise procurement channel this model reaches. The connector's availability across Copilot Studio, Logic Apps, Power Apps, and Power Automate embeds Zuva's extraction capability directly into existing enterprise automation stacks, distinct from and running in parallel with the direct Zuva Analyze product.

Developers building extraction pipelines on open-source tooling may also evaluate Unstract, which offers a no-code large language model platform with hallucination mitigation for similar structured extraction tasks.

Enterprise contract analysis

For organizations processing high volumes of business agreements, including NDAs, MSAs, and licensing contracts, Zuva's 1,300+ pre-trained AI fields provide clause extraction and classification without custom training. The Kofax Omnipage OCR engine handles degraded or scanned source documents. Cloud or on-premise Kubernetes deployment accommodates data residency requirements common in regulated enterprise environments.

Teams requiring structured extraction from financial or legal documents with source grounding may also consider LangExtract, Google's open-source Python library built for that use case. Organizations with broader contract intelligence needs spanning procurement and vendor management may find Evolution AI, which serves NatWest and Deutsche Bank on complex financial and legal document processing, a relevant alternative to evaluate alongside Zuva.

Technical specifications

Feature Specification
Supported formats 60+ document and image formats (PDF, DOC, DOCX, TXT)
OCR engine Kofax Omnipage with ML post-processing
Output formats Plain text, images, Protobuf layout files
Integration RESTful APIs, Python SDK, Power Automate, Copilot Studio, Logic Apps, Power Apps
Deployment Cloud-hosted or on-premise Kubernetes
Document retention Automatic expiration (7 to 14 days)
AI fields 1,300+ pre-trained fields for 36+ business agreement types
API throttling 100 calls per connection per 60 seconds
Max fields per request 100
Pricing Subscription and per-project options; free browser-based tier available
End-user product Zuva Analyze (GA September 2024)
Market restriction Corporate legal departments only; law firms excluded by contract
Free tier limits PDF under 150 pages or 5 MB; no export or copy/paste

Zuva's Microsoft connector excludes US Government GCC, GCC High, China Cloud, and DoD environments. Organizations with federal or sovereign cloud requirements should confirm deployment options directly with Zuva before evaluation.

Resources

Company information

Website: https://zuva.ai/ Founded: 2021 Employees: 34 Location: Toronto, Canada Parent: Independent (spun out from Kira Systems following Litera acquisition) Founders: Noah Waisberg (CEO) and Alexander Hudek (co-founders of Kira Systems) Board: Peter Sobiloff (managing director, Insight Partners) joined at Series A