Plexus uses enterprise-grade generative AI to help legal and business teams work faster, more accurately, and at scale. AI is embedded across the platform – from document analysis and contract review to intelligent recommendations and self-service legal guidance.
This article explains how AI is used on Plexus, how your data is protected, and what you can expect from AI-powered features.
Your data is secure
Plexus is built on a foundation of enterprise-grade security, and our AI features are no exception.
Your data is never used to train public AI models. Plexus uses OpenAI's enterprise-grade API to power its AI features. Under OpenAI's enterprise privacy commitments, API data is not used to train or improve public models — full stop. Your documents, conversations, and legal content remain yours.
Your data is isolated. Documents and AI context are stored in isolated databases and vector stores, ensuring zero data crossover between tenants. Customer data stays within your secure environment and is never shared or repurposed.
AI is enhanced with your organisation's expertise. Plexus AI is powered by a foundation model with broad language understanding, but its recommendations are grounded in the documents, playbooks, and policies you upload. This means outputs are tailored to your organisation's standards and risk appetite — not generic advice.
Encryption and access controls are built in. All data is encrypted in transit (TLS 1.2+) and at rest (AES-256), with strict role-based access controls and full audit logging across all AI operations.
Plexus is independently certified. We maintain ISO 27001:2022 certification and SOC 2 Type 2 compliance. Our infrastructure is hosted on AWS with firewalls, monitoring, and intrusion detection.
Top tip! For comprehensive details on our security posture, certifications, and compliance controls, visit plexus.co/security. From there, you can also register for access to our security portal — the source of truth for all security documentation and artefacts.
Accuracy and human oversight
Generative AI can produce text that sounds plausible but does not accurately reflect the source material, commonly known as "hallucination". Plexus reduces this risk through several layers of safeguards:
Grounded in your documents. Where you provide documents and Knowledge, features draw on your own material, such as playbooks and precedents, rather than general internet data. This reduces fabricated outputs and aligns results to your standards.
Without Knowledge, features fall back to the broader model, producing more generic output that warrants closer review. Note that Document Summaries and Draft Advice draw on the broader model in all cases.
Constrained by design. Prompts and guardrails restrict outputs to relevant information drawn from the provided context.
Clearly labelled. All AI-generated content is tagged, so its origin is always clear.
Human review. AI is designed to support, not replace, human decisions. The exact controls vary by feature: applying redlines or publishing draft advice requires a person to approve them first, most outputs can be edited. Review all AI output before relying or sharing it.
Intellectual property and copyright
Plexus AI generates original text using statistical language modelling, guided by the specific context of your documents and prompts. Given how the features work, it is unlikely that any substantial part of a copyrighted work would be reproduced.
Our enterprise AI provider offers intellectual property indemnity for enterprise use, subject to the terms of that agreement.
Note that text generated by AI is generally not eligible for copyright protection in most jurisdictions.
AI Features on Plexus
Plexus Counsel
Plexus Counsel is your AI-powered digital lawyer, embedded directly into Plexus workflows. It reads your documents end-to-end, interprets them through the lens of your organisation's policies, playbooks, and precedents, and provides clear, accessible guidance in plain language.
Counsel supports both business users and legal teams – from answering questions about document content and summarising context, to identifying risks and recommending next steps. It operates as a trusted digital colleague, helping lawyers focus on strategic work while routine queries are handled instantly.
Plexus Redlining
Plexus Review & Redline suggests changes to your contracts directly inside Microsoft Word. It reads your document end to end, weighs it against your organisation's playbooks and precedents, and proposes clause-level redlines to bring the contract into line with your standards and speed up review.
Every change is proposed as a pending suggestion that you review before anything is applied. You can accept, reject, or refine each one, and accepted changes appear as tracked changes so you and your colleagues can see and consolidate them exactly as you would any other edit.
With Plexus Knowledge, the redlines are recommendations grounded in your own Knowledge, with a person in control of what reaches the document.
Plexus Knowledge
Plexus Knowledge turns your legal playbooks, policies, and precedents into an AI-powered guidance engine. Rather than storing rules in disconnected folders or static documents, Knowledge embeds your organisation's expertise directly into workflows — so your Plexus AI recommendations align with how your legal team thinks about risk and commercial outcomes.
Contract Review
The AI-powered contract review feature helps legal teams accelerate their review process while maintaining accuracy and consistency. When a contract is uploaded or generated in Plexus, the AI analyses its contents and provides clause-level insights — identifying key risks, unusual terms, and potential deviations from your approved precedents.
Lawyers can review documents with or without their organisation's precedent library enabled, allowing flexibility depending on the use case. AI recommendations appear directly alongside relevant clauses, helping reviewers focus their attention where it matters most.
Document Summaries
Document summaries provide users with a concise overview of a document's contents in the Key Facts tab, so important details can be understood at a glance. The AI reads the document and prioritises key information such as counterparties, key dates, and jurisdiction.
AI summaries are clearly tagged and can be reviewed and edited by users — for example, if details change or additional context needs to be added.
AI-Generated Questions
AI-generated questions appear on the intake form of the Request Legal Support app. Generative AI analyses the text of a request, identifies gaps in the information provided, and generates targeted questions to gather all relevant facts from the business user — reducing back-and-forth between business users and the legal team.
Draft Advice
The AI-generated draft advice feature analyses the information provided through the Request Legal Support app — including uploaded documents — and prepares a draft that identifies the key legal issues and a proposed course of action. The assigned lawyer can then review and edit the draft before publishing advice in the activity feed.
What are the risks?
Plexus AI is designed to keep a person in control, so the risks are manageable as long as you treat AI output as a first draft rather than a final answer. The main things to keep in mind:
Outputs need review: AI can produce information that is inaccurate, incomplete, or not quite right for your situation. Always review and edit AI-generated content before you rely on it or send it.
A starting point, not legal advice: AI output is general guidance to speed you up, not final legal advice. Anything with legal or commercial consequences should be reviewed by a qualified person. Draft Advice, for example, is a head start for lawyers to review and edit, not finished advice.
Quality depends on your Knowledge: Plexus AI bases its answers on the documents and Knowledge your organisation uploads, such as your playbooks and precedents. If those are outdated, inconsistent, or contain sensitive material, that flows through to recommendations. Keep your Knowledge current and review what you upload.
On your data: it stays within your own isolated environment and is not accessible to other customers, and it is not used to train the underlying AI models. It is used only to answer your prompts and generate your outputs.
Learn more
Plexus is continuously developing new AI capabilities as part of our product roadmap. You can keep up to date with new features and improvements on our product release notes.
For detailed information about our security practices, certifications, and AI data handling, visit plexus.co/security.