Where AI is used, and where a person decides.
The assistant on this site runs on Claude from Anthropic. It says it is an AI, it sees only what you type, and every intake brief it writes goes to a person on the team. Inside the firm, anything sent to a client or published is reviewed by a person first.
The assistant on this site
Two modes. The intake asks about your situation and writes a brief for the team. The questions assistant answers questions about the firm and asks for no personal details.
What it sees. What you type, and for an intake, the contact details you give it.
What it never does. Quote a price, commit the firm to anything, or book a time. A team member reviews what you shared and gets in touch by email.
A person is always available. Email info@corsacvantage.com and the team replies directly.
The model provider
Anthropic receives each conversation to generate the replies. Under Anthropic's commercial terms, inputs and outputs sent through its API are not used to train its models. Retention on Anthropic's side follows those terms.
If you dictate, your browser's speech service transcribes the audio. In Chrome that service is Google's. Typing sends no audio anywhere.
Inside the firm
Corsac Vantage runs on The AI Den, its own governed agent roster with tiered authority, hard-veto roles, and escalation protocols. Agents draft and research. A person reviews anything that leaves the firm.
In client work, which AI tools may handle a client's data is agreed at scoping and written into the engagement agreement.