Shadow AI isn’t going away. Your legal policy needs to follow it.

AI adoption in legal is moving quickly, but governance is not always keeping pace. Axiom’s 2026 survey of 528 in-house legal leaders found that two-thirds of surveyed teams were using general-purpose AI in its default configuration.

The issue is not simply which AI tool people use. The bigger question is what standards those tools are working against, and who decides what is acceptable.

AI is already part of the way people work, which means the challenge for Legal is no longer simply whether to allow it, but how to maintain control as the number of tools and ways of using them continues to grow.

Blocking AI isn’t the answer

One person might use ChatGPT to help draft a clause, another might review an agreement with Claude, while someone else works with Copilot inside an existing workflow. As AI becomes embedded in more of the software people already use, trying to control every individual tool becomes increasingly difficult.

Approving a particular AI tool does not by itself solve the underlying problem. The tool still needs access to the organisation’s approved legal standards and a reliable way to apply them. It does not know which liability caps are acceptable, which governing laws Legal has approved, which clauses are mandatory, what can be negotiated or which issues need to be escalated.

These are decisions that should be governed by your organisation’s approved legal policy, rather than determined by whichever AI tool happens to be used.

The rule should follow the work

The choice of AI tool still matters. Security, confidentiality and data handling all need to be considered, but your legal standards should not change depending on which interface someone happens to use.

Whether a contract is being drafted with ChatGPT, reviewed with Claude or worked on in Word, the same approved legal policy should determine what is acceptable.

And this is not only an AI problem. The same applies to counterparty paper, contracts drafted internally and documents moving through other systems. The source of the contract should not determine the standard it is held to.

Rather than rebuilding Legal’s standards in every new tool, one approved policy can govern contract work across Lexnus and connected tools.

This is what Lexnus is built for

Lexnus turns your organisation’s approved legal positions into a policy layer that can be applied consistently across contract workflows.

Through Lexnus’s MCP integration, MCP-compatible AI tools can access the same approved playbooks, clauses and policy rules. AI can still help people draft, review and work with contracts, while the organisation’s approved policy determines what is acceptable, what requires attention and what needs to be escalated.

This means Legal does not have to recreate its standards every time the business adopts a new AI tool or changes the way it works. The tools can continue to evolve while the organisation’s legal policy remains consistent.

Shadow AI isn’t going away. Legal cannot control every new AI tool people choose to use. But it can make sure that contract work running through Lexnus and connected tools is checked against the same approved standards.