Scoring Engine | Arbitra

Logic Core

Beyond simple pass/fail.

Real conversations are nuanced. Arbitra's scoring engine supports complex logic to match how your contracts define quality.

Scoring Criteria

Add or subtract points using weighted criteria.

+10 points: Appointment booked

Rule Overrides

Force billable or non-billable when rules trigger.

setBillableTrue: tech_on_way

Required Fields

Missing required fields marks a call incomplete.

Missing outcome -> status: incomplete

LOGIC SIMULATIONSTATUS: ACTIVE

Appointment Booked

+10

Budget Confirmed

+25

No Next Step

-15

Total Score

20/ 100 possible

Interactive Demo

Build a scoring policy (and see the receipts)

A lightweight mock of Arbitra’s scoring setup. Edit a template, test it on a sample call, and see deterministic output with evidence. Saved locally in your browser.

On mobile, switch between builder and results. On desktop, both are shown side by side.

Policy Builder

Templates + config. No sign-in. No data leaves your device.

Setup

Template

Focus on lead quality: intent, qualification, solution fit, and next steps.

Policy name

Policy version

Billable threshold

Contract V2

Missing required criteria marks a call incomplete (reviewable).

Evaluation provider

Demo only. Production supports bring-your-own keys (OpenAI, Anthropic, Bedrock, and more).

Transcription provider

Criteria

Label

Weight

Evidence line

Evidence quote

Label

Weight

Evidence line

Evidence quote

Label

Weight

Evidence line

Evidence quote

Label

Weight

Evidence line

Evidence quote

Tiers

Name

Min

Max

Name

Min

Max

Name

Min

Max

Name

Min

Max

Test call controls

Call duration (seconds)

Short-call rule triggers under 30s.

Manual override

Loading saved demo…

Live output

Deterministic score, tier, billable status, and receipts.

OpenAIAssemblyAI (default)

Score

50/100

Tier: Weak Lead

Billable

Complete

Meets threshold (30+)

billable

true

statuscomplete

tierWeak Lead

policyPolicy v2

Receipts (evidence)

Only shows criteria marked “Met”.

Understands problem

Line 1

“my AC stopped working…”

Clear next steps

Line 7

“Yes — book it for 2.”

Transcript snippet

Demo transcript

252s

1CallerHey — I’m calling because my AC stopped working and it’s 95 degrees in the house.

2RepGot it. We can help. What city are you in?

3CallerAustin.

4RepOkay. Are you looking to schedule service, or just get a quote first?

5CallerIf you can come out tomorrow, I want to schedule.

6RepWe have a 2pm opening tomorrow. Want me to book it?

7CallerYes — book it for 2.

8RepGreat. What’s your address and the best number to reach you?

This is a mock UI for the marketing site. The real product supports deeper configuration, templates, review workflows, and exports—provisioned invite-only.

What this demonstrates

Deterministic scoring (same inputs, same outputs)
Receipts attached to each decision (transcript-linked)
Config over code (copyable JSON policy)

Request Private Pilot

Invite-only deployments. Reserve a Private Deployment Sprint to deploy your instance and encode your criteria.

For Agencies

Built for Multi-Client Operations

Manage scoring across multiple clients without rebuilding your logic from scratch.

Templates + Presets

Start from proven templates and tailor criteria, tiers, and rules for each client.

Per-Client Scoring

Each client gets their own scoring config without copying the entire workspace.

Client-Ready Reports

Export PDF and CSV reports with scores, evidence, and outcomes.

Private Instance

Single-tenant deployment with multi-client data separation inside your org.

Turn your contract into a scoring system

Use templates or build your own rubric. We will help you encode it.

Request Private Pilot

Precision control over billable standards

Beyond simple pass/fail.

Scoring Criteria

Rule Overrides

Required Fields

Build a scoring policy (and see the receipts)

Built for Multi-Client Operations

Templates + Presets

Per-Client Scoring

Client-Ready Reports

Private Instance

Turn your contract into a scoring system