AI scoring and rubrics · Autessa Roleplay
Criteria for every kind of skill
Autessa Roleplay criteria come in several kinds, so each part of a conversation is scored the right way: Rated on a scale between two anchors, Yes / no, Framework (a named method such as MEDDIC with its own rated parts), Knowledge check against reference text, and Feedback only, which writes notes but never moves the score.
The problem it solves
Not everything in a conversation is a one-to-five
"Did they agree a next step?" is yes or no. "Did they quote the refund policy correctly?" is a fact check. "Did they qualify the deal?" is a whole framework. Squeezing all of it onto one scale makes scores vague. Different kinds of criteria make them precise.
What you get
- Rated: scored 1 to the rubric maximum, between "what a 1 looks like" and "what a top score looks like".
- Yes / no: either they did it or they did not. No partial credit. Shown as Achieved or Not achieved.
- Framework: a named method such as MEDDIC or SPIN; its rated parts are averaged into one score.
- Knowledge check: statements checked against reference text you provide, so wrong facts are caught.
- Feedback only: written feedback that never moves the score, for things like rapport, warmth or composure.
- Mix kinds in one rubric, weighted however you like.
The run-through
How does criteria for every kind of skill work?
- SCENE 1
Pick the kind
Choose a kind for each criterion in the builder; the form shows only the fields that kind needs.
- SCENE 2
Fill in what it needs
Anchors for rated, a framework name and parts for frameworks, reference text for knowledge checks.
- SCENE 3
Weight it
Set the weight; feedback-only criteria are listed as "Not scored".
- SCENE 4
Read the result
The report shows each kind its own way: a value, Achieved, Correct or Incorrect, or Feedback.
The cue sheet
What each kind asks for and shows
| Option | What it does |
|---|---|
| Rated | Instruction to the scorer, plus what a 1 looks like and what a top score looks like. Shows a value on the scale. |
| Yes / no | Instruction to the scorer. Shows Achieved or Not achieved. |
| Framework | Framework name and named parts, each with its own instruction and anchors. Shows one averaged score and each part. |
| Knowledge check | Reference text the trainee's statements are checked against. Shows each claim as Correct, Incorrect or Not supported. |
| Feedback only | Instruction to the scorer. Shows written feedback; never changes the score. |
The cast
Who can do what?
| Who | Can |
|---|---|
| Coach | Choose kinds, write anchors and reference text, and reorder criteria. |
| Trainee | Read each criterion and its kind in the library and on reports. |
What the AI does
Behind the scenes
The scorer is given the instruction, the anchors, the framework parts or the reference text for each criterion and returns a value, a rationale and the quotes it relied on.
Questions
Questions about criteria for every kind of skill
Can Autessa Roleplay score MEDDIC?
Yes. Add a framework criterion named MEDDIC with parts for Metrics, Economic buyer, Decision criteria, Decision process, Identify pain and Champion. The starter sales rubric does exactly this.
What is a feedback-only criterion for?
For qualities you want commented on but not graded, such as rapport, warmth or composure. It writes feedback and never moves the score.
Do yes/no criteria give partial credit?
No. They are achieved or not achieved.
100% customizable, code and all
Make Autessa Roleplay fully your own.
Use it out of the box, then change anything. Ask Autobot, Autessa’s AI builder, for a new field, screen, report, workflow or agent and it builds it into your copy of the app. Or open the code and edit it yourself. Nothing is locked: the app, its data and its code are yours.
Start free and make it yours