The compliance decision record for AI calls and texts
The Standard CRM records every outbound decision immutably before it fires. Before Atlas, the AI brain, places a call or sends a text, the system writes down the reasoning behind that contact, including the consent and quiet-hours checks it ran, the A2P check it ran on any text, and the result. The decision lands in the ledger first, and only then does the contact go out. If anyone ever asks why a particular contact was made, you can produce the answer on demand.
This is the spine of how the product works. The AI plans the next move, but fixed, fixed rules decide whether it is allowed, and the ledger captures that decision so the whole system is fast and accountable at the same time.
10:24:58
Outbound call to Carol R.
Consent on file · Quiet hours OK · State allows AI calls. Status: sent.
10:25:12
Booking confirmation SMS
A2P registered · Consent on file. Status: sent.
21:40:03
Outbound SMS to Linda P.
Quiet hours: blocked (local 9:40 PM). Status: blocked.
- Before it fires
- Every decision recorded first
- Deterministic
- Rules gate, not the AI
- On demand
- Retrieve any contact’s reasoning
How does record-before-fire work?
The order is the point. Every outbound action runs the same gate, in the same sequence, and the record is written before the contact ever happens.
Step 1
Consent is confirmed
The system confirms there is consent on file for this kind of contact.
Step 2
Do-not-call holds are applied
An imported lead with a confirmed do-not-call listing is held for review.
Step 3
Quiet hours are enforced
Local quiet hours are respected, so a contact never goes out at the wrong time.
Step 4
A2P registration is verified
For texts, the sending number must be A2P registered before a message can send.
Step 5
The decision is recorded
The outcome and the checks that produced it are persisted to the permanent record.
Step 6
Only then does it fire
The call or text goes out last, so every contact already has its paper trail.
Deterministic checks
Consent and quiet hours are checked by fixed rules, and A2P for texts, never left to the AI.
Immutable record
Each decision is written down before the contact fires and cannot be quietly edited later.
On-demand retrieval
Pull the reasoning behind any single call or text whenever you need it.
What can you produce for a TCPA inquiry?
If a question arises about a specific contact, you can retrieve the decision behind it: the checks that ran, the inputs at the time, and why the contact was permitted. Instead of reconstructing what might have happened from scattered logs, you produce the recorded decision as it was made. The system is mechanism-first by design, so the explanation already exists before you ever need it.
Why does the AI plan but the rules gate?
Letting a model decide on its own whether a contact is legal would be unpredictable. Instead, Atlas proposes the next action, and a fixed-rule layer checks consent and quiet hours against fixed logic, and A2P registration before every text. The rules have the final say, every time, and their verdict is what gets recorded. That separation is what keeps behavior consistent and auditable rather than improvised call to call.
Frequently asked questions
What is the compliance ledger?
It is a permanent record of every outbound decision the system makes. Before any call or text fires, the reason it is allowed, including the consent and quiet-hours checks, and the A2P check on texts, is written down. Nothing goes out without a recorded decision behind it.
When is a decision recorded, before or after the contact?
Before. The decision is persisted first, then the contact fires. This ordering is deliberate: you never end up with a call that happened but was never accounted for.
Can I produce a record for a specific contact?
Yes. You can retrieve the decision behind any individual call or text on demand, with the checks that were run and the inputs at the time. If a question ever comes up about why a contact was made, the answer is on file.
How are the compliance checks decided?
Deterministically. The AI plans the next action, but fixed rules check consent and quiet hours, and A2P registration before every text, and decide whether the contact is allowed. The rules gate; the AI does not get to override them.
Why can the records never be changed?
An audit trail is only useful if it cannot be quietly edited after the fact. Recording each decision immutably means the history you produce later is the history as it actually happened.
See plans and pricing.
Every plan runs the full platform and starts with a 7-day free trial.
See plans and pricing