Governance AI
AI Policy.
AI for speed, qualified people for judgment, and a documented review behind every deliverable.
Lex Auxilium was built on a single conviction: the future of professional support belongs to teams that combine the speed of artificial intelligence with the judgment of qualified people. This policy explains how we use AI in our work, the safeguards around it, and the standards our clients can hold us to.
01Our position on AI
We treat AI as a force multiplier, never as a substitute for professional judgment. AI helps us draft, research and reconcile faster than a manual process ever could. It does not decide, advise or sign off. Every output that leaves us has been reviewed by a qualified person who is accountable for it.
02How we use AI in our work
Across our services, we use AI to accelerate well-defined tasks, including:
- Preparing first drafts of documents, summaries and correspondence for review.
- Speeding up legal and factual research, with sources verified by hand.
- Reconciling records and surfacing anomalies in accounting and bookkeeping work.
- Supporting AI legal data annotation, such as clause tagging, entity recognition, case-outcome classification, citation linking and review of model outputs.
In every case, AI assists a qualified person. It does not work unsupervised, and it does not produce client deliverables on its own.
03Human verification and accountability
Every deliverable passes a documented two-stage review. Work is produced with AI assistance, then checked by a qualified reviewer, with citations and key facts verified against primary sources. Each client has a named engagement lead who is responsible for the quality of the work. The result is the standard we describe as lawyer-grade reasoning at paralegal-tier rates: fast to produce, and checked by people who stand behind it.
04Confidentiality and client data
Protecting client information is a precondition for how we use AI, not an afterthought. Our commitments include:
- We do not allow confidential client information to be used to train public, third-party AI models.
- We use approved tools and controlled configurations for any work that involves client material.
- Access to client information is limited to the people assigned to the engagement, each bound by confidentiality obligations.
- Where a client has specific restrictions on AI use, we follow them.
05Governance and security
We maintain a written AI governance framework that sits alongside our confidentiality and security practices. It covers which tools may be used, how client data may and may not be handled, and the review steps that every deliverable must pass. We review this framework as the technology and the rules around it continue to change.
06Accuracy and limitations
AI systems can produce confident but incorrect results. That is precisely why human verification is built into everything we do. We do not rely on unverified AI output, and we correct and stand behind the verified deliverables we provide. If we identify an error in our work, we will address it promptly.
07Transparency and client control
We are happy to explain how AI is used on a given engagement, and clients can set preferences or limits on its use. If you would like more detail about the tools or controls that apply to your work, please ask your engagement lead.
08Professional and jurisdictional responsibility
Our advocates work in a paralegal and support capacity under the supervision of the client firm's attorneys, who remain responsible for advising their clients and for compliance with the professional rules of their jurisdiction. Our use of AI is designed to support that responsibility, never to bypass it.
09Updates to this policy
We may update this policy as our practices and the surrounding technology evolve. When we do, we will revise the date shown above.
10Contact us
If you have questions about how we use AI, please contact us at hello@lexauxilium.com.