How We Use AI
We use AI to speed up research and first drafts. It does not decide what is true about your environment, what a risk is worth to your business, or what you should approve. Those decisions stay with people, and our use of AI is governed by a written policy.
Where AI helps, and where it stops
Policies and procedures. AI proposes structure and wording. We rewrite it to match how your team actually works.
Standards and guidance. AI helps summarize published frameworks. We decide what fits your size, risk and obligations, and check summaries against the source.
Security questionnaires. AI can suggest answers from your approved answer library. We check them against your current controls, under an agreed review protocol, because an answer that was true at your last audit may not be true today.
Tabletop exercises. AI helps with scenario detail. We choose the scenario based on your real systems and dependencies.
What stays with people
Context. Your systems, your history, and the commitments you have made to customers and auditors.
Verification. AI can be confident and wrong. Facts, statistics and regulatory references are checked against primary sources before we rely on them.
Judgment. What to prioritize, what a risk is worth, and what to recommend.
Accountability. AI does not make decisions or approve deliverables. Every deliverable is reviewed and owned by BARE's principal.
Our guardrails
Vetted tools only. Any tool that handles client data is reviewed first, including its data processing terms, where data is processed, and whether it trains on inputs.
Your data does not train AI models. Our assistant for client work runs on a business account whose terms prohibit training on customer content.
No AI on your systems. AI tools never act on client systems, never hold client credentials, and never run in the background on client data.
Disclosure. Deliverables where AI produced substantial parts of the content are marked as such, and reviewed before they reach you.
AI changes how quickly a first draft appears. It does not change who is responsible for getting it right.
Questions about how this applies to your engagement? Get in touch.

