The Otto platform

The rules engine — blocking, not logging

Otto's rules engine evaluates a written rule at the moment someone tries to act — place a call, advance a case to suit — and refuses the action when the rule fails, naming the rule, its recorded source, and when the action becomes permissible. The refusal itself is written to the audit trail.

What is the difference between blocking and logging?

Timing, and nothing else. A logging system evaluates the rule after the action, in a report someone reads later; the consumer has already had the experience by then and the remedy is remediation. A blocking system evaluates the same rule before the action and does not let it happen. The second difference matters more than it sounds: a refusal is affirmative evidence. A system that only logs can show you the violations it found. A system that blocks can show you the attempts it stopped, with the rule and the timestamp attached.

What does Otto check before a call is placed?

Five things, in order, and the first failure ends it. One: is there a collection hold on the account — a dispute, a cease-communication request, an attorney-represented flag, a bankruptcy stay, or an SCRA flag. Two: does the client prohibit outbound dialing on this portfolio at all, which some do contractually. Three: has the initial validation demand actually gone out, because no collection contact is permitted on this account before it has. Four: what time is it where the consumer is, computed from the consumer's own time zone rather than the collector's. Five: how many attempts have already been made inside the rolling window, and has there been a telephone conversation that started a cooldown. Each failure names the rule and the source recorded against it, and the source strings in the shipped configuration point at Reg F, the FDCPA, the bankruptcy code, the SCRA and the Florida overlay. TODO: verify each source string against the primary text before publication, and confirm the quiet-hours basis for every configured state rather than relying on the federal floor.

What blocks a suit from being filed?

A separate gate with two conditions, both from the client's own rule: the account-level media the client requires must be attached, and an attorney review event must exist on the file. In the shipped configuration the required media is chain of title, account statements and the signed application, and the review requirement is described as a meaningful-involvement condition. When the gate fails, the filing action does not advance and the reasons are named individually — which document is missing, or that no review is on file. The block is written as a rule-gate-blocked event carrying those reasons. TODO: verify the consent-order citation recorded as the source of this rule, and check whether the media list should be stated as a client requirement rather than as a regulatory one.

Can a block be overridden?

One of them, by two roles. The contact-frequency cap can be overridden by a compliance officer or an administrator, and doing so requires a typed justification which is written to the audit trail along with the name of the person who gave it and the fact that the subsequent call was overridden. Nothing else is overridable in the product. A dispute hold, a cease request, an attorney-representation flag, a bankruptcy stay, an SCRA flag, a client's outbound prohibition, the demand-first gate and quiet hours have no override path at all — not a harder one, none.

What runs on a schedule rather than on an action?

A control calendar, because some checks are not triggered by anyone doing anything. Bankruptcy scrubs run daily against the docket sources, the SCRA scrub runs weekly, the deceased scrub runs weekly, and a state license and bond check runs monthly across the configured jurisdictions. Each run is written to the audit trail with what was checked, what was flagged and what the system did about it — and a flag becomes a hold without a human step, which is the point. A manual run is attributed to the person who started it.

How much of the engine is built?

The evaluation logic is real and it is small enough to read. Contact-frequency evaluation counts attempts in the rolling window and applies the post-conversation cooldown; the suit gate compares required media against attached media and looks for a review event; the call path runs the five checks in sequence and returns a structured refusal. What runs against it is seeded data and a fixed clock, so a timing figure quoted from a demo screen is a seeded value and not a measurement. The rules themselves are configuration, so what a real deployment enforces is whatever was written into it.

What this does not do

Enforcement is not a compliance programme. It cannot exercise attorney judgment, it cannot tell you whether a filing is well-advised, and it cannot supervise the people using it. It applies rules that have been written down correctly, which makes writing them down correctly the work rather than remembering them. It also cannot enforce a rule for a jurisdiction nobody has configured: four states ship configured, and an account in a fifth is governed by whatever general rules apply to it until somebody adds the overlay.

This page describes Otto, not the law, and is not legal advice. Otto applies rules that have been written down correctly; it does not warrant that using it results in compliance with any statute, regulation or contract. Where a rule is named here, confirm it against the primary source and your own counsel before relying on it.