Responsible AI
AI Notice
Last updated · 15 April 2026
This notice explains where and how LightMountain Advisory (“LMA”) uses artificial intelligence on this website and in our internal workflow, what you should expect from AI-generated content, and the human checks that sit behind it.
Where AI is used
- Zayd— the public-facing chat assistant at the bottom-right of every page. Zayd answers general questions about LMA’s services, routes you to the right team member, and helps schedule meetings. Zayd is powered by a large language model hosted by Anthropic PBC.
- LMA Intelligence (the internal platform at /intelligence)— a suite of tools used by LMA partners and authorised team members to accelerate deal-origination work. It includes deal triage, teaser drafting, lender shortlisting, restructuring playbook generation, and a pipeline tracker. Its outputs are AI-assisted working drafts for internal review.
- Insights preview— a single-question preview of the LMA Intelligence tone appears on the public home page. It is restricted and rate-limited.
What AI-generated content is — and isn’t
The large language models behind Zayd and LMA Intelligence are trained on broad information about finance, regulation, and the MENA/GCC market. LMA has carefully framed these systems to respond in a senior-advisor voice and to stay within its areas of competence. Despite this, you should treat all AI-generated output as preliminary and subject to human review.
AI may:
- Name a regulation, rule, or circular incorrectly;
- Misstate lender appetite, market pricing, or the availability of a structure;
- Produce information that was accurate at training time but is no longer current;
- Generate plausible-sounding specifics that are, in fact, imprecise (“hallucinations”);
- Omit material nuance that a real file would require.
Human review
No AI-generated output is sent externally as a final LMA deliverable without partner review. Teasers, lender lists, and restructuring memos produced in LMA Intelligence are treated as first drafts. An LMA partner is responsible for reviewing, revising, and approving any material before it is shared with a client or a counterparty.
Meetings booked or enquiries submitted through Zayd are received and actioned by a human. Zayd does not finalise engagements, sign documents, or quote fees. Where Zayd flags a message as substantive, an LMA team member follows up.
Privacy
Inputs you send to Zayd or the LMA Intelligence platform are transmitted to our AI provider to produce a response. Under our agreement with the provider, those inputs are not used to train the provider’s models. LMA does not persist chat transcripts or deal inputs on its own servers beyond the ephemeral processing needed to return a response. Pipeline tracker entries are kept only in your own browser’s local storage and never leave your device.
Please do not enter highly sensitive personal data, confidential client data, or legally privileged material into AI features. Our Privacy Policy provides the full detail.
Limits and safety
Our AI features are designed to stay within LMA’s professional scope. Zayd will steer off-topic conversations back to LMA’s remit and will decline to provide specific legal, tax, or regulated financial advice. Rate limits and content safeguards apply to all AI endpoints.
Accountability
Ultimate responsibility for the quality, accuracy, and appropriateness of any advice provided by LMA rests with LMA’s partners — not with any software tool. AI accelerates the firm; it does not replace it.
Questions and corrections
If you see AI-generated content on this website that appears inaccurate or inappropriate, or you have a question about how these tools work, please contact us at info@lightmountainadvisory.com. We review these reports carefully.
Questions about this policy? Write to info@lightmountainadvisory.com.
Privacy · Terms · Disclaimer · AI Notice