Six operating fields
Company description, mission, product or service, directives (binding), escalation contact, context notes. Each field is char-limited so your token budget stays bounded forever (~3500 chars total, ~900 tokens).
An AI agent is only as good as the brief it runs on. Each agent in CloseHunt ties to an Identity · the operating layer that defines who it represents, what it can say, when to escalate, and what background facts to lean on. Reusable across agents, editable from anywhere, auto-filled from your website in seconds.
Company description, mission, product or service, directives (binding), escalation contact, context notes. Each field is char-limited so your token budget stays bounded forever (~3500 chars total, ~900 tokens).
Directives aren't suggestions, they're injected verbatim as scope rules into every reply prompt. "Never discuss pricing." "Always offer a discovery call before a demo." Done · the AI obeys.
Paste a company URL. The AI fetches the page, extracts mission, product line, brand description, and suggests directives. Validate field by field. No fabrication · better empty than guessed.
Connect an inbox under an identity. The agent only sends from that inbox, never accidentally from someone else's account. Multi-rep workspaces stay clean.
Outbound + inbound + a NURTURE campaign can all share the same voice. Update the identity once · every connected agent sees the change on its next send.
Use a real teammate's identity (with their inbox) or a workspace persona (fictional account exec). Both inject the same operating layer into every prompt.
Because the prompt has to be re-stitched at every send · with the prospect, the channel, the conversation history. Identity stores the parts that NEVER change (your brand, your scope, your escalation) so the prompt-stitching layer can stay clean and focused on the per-prospect parts.
Submit any public company URL. The analyzer fetches the page (max 6000 chars after HTML strip), sends it to an admin-configurable LLM, and returns suggested values for company name, description, product or service, mission and operating directives. You apply field-by-field · never auto-overwrite.
The agent's identity directives are strict · they block any clever phrasing the prospect tries. Combined with your negative-reply policy and your turn cap, you have a clean kill-switch for any conversation.
Every Identity field has a strict character limit (200-1500 depending on the field), enforced both client-side (visible counter) and server-side (zod schema). Total identity budget is ~3500 chars ≈ 900 tokens. It never grows.
Yes. Identity is the source of truth · edit once, every connected agent sees the change on its next send. No re-deploy, no copy-paste across agents.
Trial on request · contact our sales team.