Plans
Three software subscription plans, billed per agent seat. The line between them is how much of the work the AI does — everything involved in running a quality program by hand is in all three.
From 10 agents
The smallest contract is 10 agent seats. Evaluators and managers are included on top of that, so you are not paying for the people doing the reviewing.
Pay annually, get two months
An annual contract costs the equivalent of ten months. Same plan, same seats — you just stop paying for two of them.
Essential
A complete manual quality program, without AI.
USD, billed monthly · annual saves two months
- Unlimited manual evaluations
- Weighted scorecards with critical questions
- Agent disputes with a full audit trail
- Dashboards, insights and CSV/PDF export
- Calibration between evaluators
- Teams, supervisors and role-based access
- Masking of card numbers and identifiers
Smart AI
The AI transcribes and scores; your team reviews and signs.
USD, billed monthly · annual saves two months
- Everything in Essential
- AI transcription with speaker labels
- 500 transcription minutes per seat per month
- AI auto-scoring with per-question guidance
- 50 AI analyses per seat per month
- Call metrics, hold time, dead air and voice analysis
- Calibrate the AI against your evaluators
Elite AI
Adds your own context and metrics, and turns scores into coaching.
USD, billed monthly · annual saves two months
- Everything in Smart AI
- 2,000 transcription minutes per seat per month
- 200 AI analyses per seat per month
- Your own AI metrics reported on every conversation
- Your own context fields feeding each analysis
- Per-agent and team coaching, with PDF and email
- Conversation API and department isolation
What each plan includes
| Feature | Essential | Smart AI | Elite AI |
|---|---|---|---|
Manual evaluations Unlimited, on any channel | |||
Weighted scorecards Sections, half points and N/A | |||
Critical questions A miss zeroes its section | |||
Calls, chats, email and messaging Four channel types, each measured on what it can actually be measured on | |||
Chat and email formatter Deterministic clean-up of platform exports, previewed before it replaces anything | |||
Agent disputes With resolution and audit trail | |||
Acknowledgement and kudos The agent accepts a score, and good work is singled out with a note | |||
Dashboards Trend, agents, teams and channels | |||
Insights Most-failed questions, coverage, evaluator consistency | |||
CSV and PDF export Grouped however you need it | |||
Calibration Blind scoring and comparison report | |||
Teams and supervisors Several supervisors per team, and a default team nobody falls out of | |||
Roles and permissions Agent, evaluator, manager, account admin — enforced on every route | |||
Activity log Everything that happened, linked | |||
In-app and email notifications Evaluations, disputes, invitations and coaching, in the recipient's language | |||
Bulk user import CSV with preview and per-row checks | |||
Masking before storage Cards checked with Luhn, IBANs with mod-97, removed before anything is stored | |||
Locked evidence Once a conversation has been scored, its audio and transcript cannot be replaced | |||
Passkeys, Google sign-in and two-factor Optional per person, with backup codes | |||
English and Spanish Interface, emails and reports | |||
AI transcription Speakers, language hints, timestamps | |||
Transcription minutes Included every month with each agent seat | 500 / seat / mo | 2,000 / seat / mo | |
AI auto-scoring Drafted by the model, reviewed by a person | |||
AI analyses Included every month with each agent seat; top-up packs never expire | 50 / seat / mo | 200 / seat / mo | |
Custom AI instructions Your own rubric on every question | |||
Unverified flags The model says when the transcript did not prove it | |||
Sentiment on every analysis One scale across the platform, so a year of data stays comparable | |||
Call metrics Talk ratio, pace, filler words | |||
Hold and dead air Measured from the audio, with timestamps | |||
Voice and delivery analysis Pitch, jitter, harmonics, monotony and overlap, measured on our own service | |||
Calibrate the AI Score the model against your auditors | |||
Top-up packs Buy extra analyses or minutes; what you buy does not expire at the renewal date | |||
AI custom fields Ask the model your own questions about every conversation and get a structured answer back — outcome, product mentioned, promise made. | |||
Context fields Feed the model what it cannot know from the audio — the campaign, the offer, the policy in force that week — so it scores against the right rules. | |||
Conversation API Push conversations in, pull scores out — Elite | |||
Department isolation An account admin who sees their own department and nothing beyond it | |||
Coaching reports Per agent, grounded in real calls | |||
Team coaching view Who needs the conversation this week | |||
Coaching PDF and email Sent to the agent in their language |
Mira scores is cloud software used in your browser: nothing to install and nothing shipped. Prices are in USD, per agent seat; applicable taxes are calculated at checkout. Running a large operation or want an annual contract? Tell us your size and we will put a proposal together.
Talk to us