Every conversation reviewed, not the five somebody had time for.

Spellit Actions reviews every call your team sends in, so call quality is judged across the whole team instead of the few calls a supervisor had time for. Each conversation is scored against the same rubric, and every rep is measured by the same standard.

Calls that fall outside the job, such as an off-target inquiry, are marked as not relevant. They stay out of the success score instead of dragging it down, and a switch lets a manager look at all calls or only the relevant ones. The calls come in as an export of recordings you already make, in a format agreed in advance. Spellit Actions does not join calls or record them. Results are calculated in weekly batches after the calls happen, not during the conversation.

Yes, Spellit Actions builds the call quality rubric from the review form your team already uses, written in your own wording. The Spellit team turns that form into a list of scored items for your project.

Each item is marked as mandatory or optional, and some can be flagged as outside the rep's control so they do not count against the person. A call's score is the share of mandatory items the rep passed. The weights behind the success score and the color thresholds live in your project settings, not in code. Anyone on the team can open the full item list, including the exact prompt text the AI uses to score each item. If you change a threshold later, past weeks keep their colors: a March report still reads the way it did in March.

In Spellit Actions every call quality score opens into the sentence it was based on and the minute it was said. From a team number you go to the calls behind it, then to one call card, then to a quote with a timestamp. Clicking the quote opens the transcript scrolled to that line.

Each scored item on the call card shows the AI's reasoning next to the quote, so a reviewer reads the evidence instead of relying on an impression. A block with no data behind it is hidden rather than filled with zeros, and a missing score shows as a dash, not as 0%. The point is that a reviewer can settle a question about a score in a minute, using the words from the call.

Yes, any rep can appeal a Spellit Actions score, and a person makes the final decision. The appeal names the call and, if needed, the specific rubric item, states the score the rep thinks is right and must include a written justification. Screenshots or a video can be attached.

The person deciding sees the AI's verdict, its reasoning and the quote on one side, and the rep's argument on the other. They either apply the correction or keep the AI score, and the system will not close the appeal until they write down why. An administrator in your organization decides, and a case can be escalated to Spellit support. Every step goes into an audit log. Managers also see which rubric items get disputed most often, which usually means the item's wording needs work.

Spellit Actions delivers call quality results once a week, after that week's calls have been processed. Scores, the team board and each rep's tasks are recalculated together when the week closes, and the week selector opens on the latest week available.

The page describes the first quarter in three steps. In week one, every conversation is read. By month one, the scoring reflects your process, your wording and each stage of the funnel. By quarter one, disputes go to a person and the outcomes are on record, which is when teams start to trust the score. If the written findings for a week are not ready yet, the screen says they will appear after the next recalculation rather than showing placeholder text.

In Spellit Actions a weak call quality result becomes a task for a named rep instead of a number in a spreadsheet. First, the rubric items are ranked by how much missing each one lowers the chance of reaching a next step, multiplied by how often the team misses it. The item at the top arrives with the people who missed it and the calls behind each name, so the review meeting has a single subject.

For each rep, a rule then picks one to three tasks for the week. A task carries the moment from the rep's own call and the wording to use next time. The rep marks it done, and a manager confirms it or sends it back with a reason. Reviewers spend their hours on the calls that need judgment.