Use Cases

Using AI to Draft Contracts, with a Policy Layer

AI drafts fast. The problem is that the output is new language your organisation has never approved. The liability cap in that draft may not match your policy. The governing law clause may not reflect your standard position. The indemnity language may be something your GC has never seen. It sounds right. It may not be right for you.

The problem

Legal teams are using AI for drafting. They are drafting faster than ever. But speed without governance creates a new risk: AI-generated language that no qualified lawyer has reviewed is reaching counterparties.

General-purpose AI models are not aware of your organisation's approved clause language. They do not know your GC approved a specific liability cap last quarter. They do not know your standard termination notice period. They generate language that is legally plausible and often wrong for your specific standards.

The result: lawyers are manually checking every AI-generated draft against a policy that lives in a separate document, an email thread, or institutional memory. The time saved in drafting is spent in verification.

How Lexnus solves it

1

Connect Lexnus to your AI model via MCP

Lexnus exposes its policy engine and clause library as an MCP server. Once connected, the AI model can call Lexnus during a drafting session without the lawyer switching tools or applications.

When a lawyer asks the AI model to draft a liability clause, it calls Lexnus's lexnus_get_approved_clause tool. Lexnus returns the organisation's approved liability clause for the specific contract type, deal value, and governing law, selected automatically based on the context. The AI model uses that clause. The output is not generated from scratch. It comes from your governed library.

2

Evaluate as you draft

As the draft develops, the AI model can call lexnus_evaluate_clause on any provision it has written or that the counterparty has proposed. Lexnus evaluates it against your playbook rules and returns: pass, violation, or gap, with the specific rule triggered and the approved alternative.

This happens inline, during the drafting session. The lawyer does not need to export the document, upload it separately, or run a separate review step.

3

Run a full compliance check before sending

Before the draft leaves, the complete document is submitted to lexnus_analyse_document. Lexnus runs a full analysis against your active playbook. Every rule is checked. Every gap is identified. The lawyer reviews the flagged items, not the full contract, and the compliant draft goes out.

What this changes

Before Lexnus: AI drafts fast. The lawyer manually verifies each clause against a policy that lives in a separate document, an email thread, or institutional memory. Some things get missed. The same issues recur.

With Lexnus: The AI model drafts from your approved clause library. Clause evaluation runs inline. Full compliance check before send. The lawyer focuses on the exceptions and edge cases that actually require judgement.

The audit trail shows every clause the AI model retrieved, every call made to Lexnus, and every compliance result. The GC has visibility into what AI is doing with the organisation's contracts.

Who this is for

In-house lawyers and associates who are already using AI for drafting and want to stop spending time on manual policy verification.

General Counsel who want AI drafting to happen under governance, not in a shadow workflow they cannot see or control.

Legal Ops teams building standardised contract workflows who want AI to do the heavy lifting without giving up policy control.

Frequently asked questions

Does Lexnus restrict what the AI model can do?

No. The AI model retains full drafting capability. Lexnus adds a governed layer of approved language and policy rules that the AI model can draw from. When the AI model has access to your Lexnus library, it uses your approved clauses by default. You can still override, edit, and negotiate. Lexnus tracks the deviation and flags it for review.

What if the AI model generates a clause that is not in our library?

If the AI model generates language not found in the Lexnus library, the clause evaluation step flags it. The lawyer sees that the proposed language is unrecognised and is shown the closest approved alternative. The decision to use the generated clause or the approved one stays with the lawyer.

Does it need to be a specific AI tool?

No. Any MCP-compatible AI interface can connect to Lexnus. Any agentic pipeline that supports MCP tools will work. The connection is not dependent on the specific interface.

Does this work for counterparty paper, or only for our own drafts?

Both. Lexnus analyses any contract against your playbook, whether you drafted it, the AI model drafted it, or the counterparty sent it. The analysis pipeline is the same regardless of the document's origin.

What is the setup process?

Connect Lexnus to your AI model via the MCP server configuration. Ensure your clause library and playbooks are published in Lexnus. The connection is live immediately. No integration project. No implementation fee. Onboarding is included in all Lexnus plans.

Free tier available. No card required. Generally available 31 August 2026.