Not promises. Mechanisms.
Most AI sales tools describe safety in adjectives. This page describes ours in mechanisms — each one enforced in code, each one visible in the product. If a claim on this page ever stops being true, the product stops working that way too, loudly.
What leaves the building — and what can't
You approve. You send.
Adam researches and prepares each opportunity and its draft. Nothing is sent without your approval, and Adam does not send for you: the draft opens in your own mailbox (Gmail, Outlook or your desktop mail app) and leaves only when you press send. Adam does not send from a shared mailbox.
Enforced server-side: approving a card queues nothing, and the API refuses both sending from a shared address and autonomous mode. The button in the dashboard has no say in it.
Every address shows how sure we are
Each email address on a card carries its state: verified (a data provider confirmed the address exists), likely (a provider found it, without full verification) or unverified (found in a public source and not checked). An email recommendation is never built on an address marked invalid.
The brief validator rejects an email action on an address marked invalid, before the card reaches your screen. There is no override flag.
Your daily number is a ceiling
Adam brings at most your account's daily number of new opportunities. An extra one is held back and labeled, never squeezed in. The same organization is not suggested twice within 30 days: a new trigger is added to the earlier card while you have not decided on it. An opportunity is also held back when the organization is on your do-not-contact list, and a cold one when someone there is already in the contacts you imported.
The ceiling and these checks run where an opportunity enters the system, not in the agent's instructions. An agent cannot talk its way past them.
LinkedIn is never automated
Adam drafts connection notes and comments; a human sends them from their own account. Your LinkedIn account is not ours to risk. Same for WhatsApp — drafts only.
Evidence discipline
A fact without a source is rejected
Every factual claim in a brief must carry a URL and a date. A brief with an unsourced fact fails validation and never reaches your screen. Facts and inferences are typed separately and never mix.
claims[0] is a FACT without evidence (url+date) → HTTP 422
Published numbers carry denominators
Every performance figure we publish includes numerator, denominator and period, from our own production system. We consider an undated, denominator-free "80% improvement" a red flag — in anyone's marketing, including ours.
Your data
Tenant isolation, enforced by the database
Client data lives in Postgres with row-level security enabled and zero client-side grants: a browser session cannot read the database directly at all — every read passes through the API, which resolves your identity server-side.
No cross-client learning
What Adam learns from your approvals, edits and ratings calibrates your hunt only. Nothing about your pipeline, contacts or drafts feeds any other client.
A replayable audit trail
Every decision, edit, held-back opportunity and mode change is timestamped with an actor and a reason. Any morning can be reconstructed event by event, including the opportunities that were held back, and why.
Questions about any mechanism on this page: aviel@agentsela.com. Built by Agent Sela · Israel. · עברית