Compass · Feedback

A coach for every learner, every session

Real-time AI coaching scores performance and gives specific, actionable feedback — so people don't just complete training, they measurably get better at the job.

AI coaching

Feedback is where learning actually happens

Practice only builds skill when it's paired with good coaching — and good coaching has always been the bottleneck. There are never enough managers with enough hours to review everyone.

Compass removes that limit. After every roleplay, quiz and assessment, it analyses what the learner did and returns a clear score plus concrete, prioritised guidance on how to improve — instantly, for everyone.

AC
Coach Ava
AI Travel Coach · post-roleplay debrief
What went well
Strong discovery — you matched the twin-centre idea to her "relax + explore" split before recommending. That's exactly the move this scenario rewards.
One thing to fix
When she asked "is that the final price?", your answer was a little vague. Next time, confirm it's all-in — transfers, taxes and ATOL protection.
Got it — so reassure her it's the final, fully-protected price?
Exactly. Try the same scenario again and lead with that — I'll watch for it.
What the coach does

Specific, scored, and always actionable

Not a vague thumbs-up — concrete guidance a learner can act on in their next attempt.

Performance scoring

Every session is graded against the skills that matter for the scenario, giving a clear, comparable score.

Real-time delivery

Feedback arrives the moment a session ends, while the experience is still fresh and easy to correct.

Strengths & gaps

The coach calls out what went well to reinforce it, and pinpoints the one or two things to fix next.

Progress over time

Scores trend across attempts so learners and managers can see skill genuinely improving.

Coaching at scale

Every employee gets the same quality of review — no waiting for a manager's calendar to free up.

AI-graded assessments

Dynamic quizzes are graded automatically with feedback, and completion earns a PDF certificate.

How it's scored

Know exactly which skill to fix next

A structured breakdown tells each learner precisely where to focus — scored 0–10 across four dimensions, every time.

Product Knowledge

Were the facts right and grounded in your material?

Needs Analysis

Did they ask the right questions and listen?

Objection Handling

Did they meet pushback with confidence?

Closing & Coverage

Did they advance it and cover every point?

Each dimension carries its own reasoning and comment, plus an overall performance category — from EXEMPLARY and ACCURATE_HELPFUL to flags like HALLUCINATION — with passmark gating so only genuinely competent learners certify.

See a sample scorecard
Good effort
7
/10
Overall score · pass mark 6.0
Skill breakdown
Product Knowledge 8/10
Strong on the twin-centre option and what the all-inclusive covers.
Needs Analysis 9/10
Good discovery on budget flexibility and dates before recommending.
Objection Handling 6/10
Quoted the package price but didn't pre-empt the "is that final?" worry.
Closing & Coverage 7/10
Offered a hold to advance the booking; didn't cover the cancellation worry.
Examiner's notes — why you got this score
A genuinely strong attempt. You matched the twin-centre idea to her "relax + explore" split really well, and your discovery on budget and dates was the standout. To push into the top band, pre-empt the "is that the final price?" worry before she raises it, and reassure her on cancellation cover.
Discuss this attempt with your coach Try again
Dedicated AI coaches

A coach that debriefs, not just grades

Compass includes dedicated coach AI entities that run post-roleplay debrief sessions — talking the learner through what happened, reinforcing strengths and prioritising the one or two fixes that matter most.

Because a coach is a template, one well-designed coach scales to unlimited learners with identical quality and consistency.

Why a buyer cares: you get the depth of a 1:1 debrief at the cost and scale of software — no waiting for a manager's calendar.

Skill breakdown · this learner
Product Knowledge8 / 10
Needs Analysis9 / 10
Objection Handling6 / 10
Closing & Coverage7 / 10
For L&D admins

Turn feedback into a measurable programme

Individual coaching rolls up into advanced analytics. Admins see progress, engagement, completion and roleplay performance across teams — spotting skills gaps before they cost a deal or a customer.

It's the proof L&D leaders have always wanted: a direct line from training activity to measurable capability.

Coaching that scales without limits

One coach template, unlimited learners, the same quality of feedback for everyone.

Book a demo
4
Scored dimensions
Instant
Feedback delivery
1→∞
Template to learners
24/7
Always available
FAQ

Coaching & feedback questions

What exactly does each dimension score mean?

Every roleplay is graded 0–10 on Product Knowledge, Needs Analysis, Objection Handling and Closing & Coverage — each with its own written reasoning, so a learner sees precisely which skill let them down rather than a single opaque number.

Can the AI penalise made-up answers?

Yes. Performance categories include flags such as HALLUCINATION, so a confident but factually wrong answer is caught and scored down — exactly the behaviour you want to coach out before it reaches a customer.

Does feedback replace human managers?

No — it frees them. AI handles the high-volume, repeatable review and debrief, so managers spend their limited time on the judgement calls and edge cases where they add the most value.

Give every learner a coach

See how Compass's real-time feedback scores performance and drives measurable improvement at scale.