skyNote Coach
Eleven screens covering FR-1 to FR-16 of the functional requirements: the private personal coach, the sales scorecard, the manager's coaching workflow, administration and mobile.
skyNote already listens to every meeting. Coach turns that recording into private, measured feedback, scored against a healthy range and backed by the exact seconds of the call where it happened. For sales teams the same engine grades calls against a rubric the team defines, so a manager can coach ten reps in the time it used to take to review one.
Every number is one click from the seconds of audio that produced it. A metric with no moment behind it is an opinion, and Coach never shows one.
The signature component is a bar carrying four marks: the ideal band, your value, your 30 day baseline and, on team screens, the top performer band. Status is written in words as well as colour.
Personal Coach carries a lock and a neutral callout. Team data carries the blue users callout. The two are never mixed on one card, and the disclosure appears before the first metric exists.
Fifteen metrics is a wall. The misses sort to the top, advice stops at three and goals stop at three. Everything else is available, nothing else is pushed.
Every failure state names its cause. Every AI grade carries one sentence of justification and the moments behind it, and a human override that wins.
No composite rep score, no ranking without an explicit metric, no HR surface. The manager's page states what it cannot show.
Every recorded meeting. Coach status shows whether a private score exists for you on that meeting.
Aug 25, 2026. 1h 31m. 9 people, 7 voices. You are matched to the voice labelled Victor Reyes.
Identity resolution, the in progress state, the cannot run state and the low signal state.
Every moment behind one number, ordered by how much each contributed.
Your digital twin. She watches your meetings, your calendar and your open items, and proposes the next move. Nothing leaves your account until you approve it.
One click turns a meeting into a ready follow-up. It goes out through your own Outlook, Gmail or Teams account.
She answers what you say. End the session when you are done.
How the whole conversation went, against the criteria for this stage.
Your score and pillars over time, one metric in depth, the goals you are working on, and a place to rehearse before the real conversation. Private to you.
The prospects your CRM knows about, and the coaching that context makes possible. This destination exists because the workspace has a CRM connected.
28 clips, built on soundbites, so every clip already has a share link and an AI synthesis.
Skypoint workspace, 482 users, Coach enabled 14 Aug 2026. Admins configure Coach. No admin role can read anyone's Coach data.
The same information on a phone, one tap in person capture, and the weekly digest that links straight into a moment.
The moment a person cares how they spoke is the moment they open the call. A separate Coach product gets visited on launch day and never again. My Coach exists for trends. The meeting tab is where the habit forms.
My Coach answers how you speak, which is the same question for a counselor and an engineer, so it never changes shape. My Discovery answers who you are speaking to, which only has an answer when a CRM is connected. Splitting on that line means nobody watches half of a screen disappear when an integration is switched off, and the CRM state is one attribute on the root element driving the nav item, the prospect column and the meeting panel together.
Sub screens are reached from their parent and highlight the parent nav item. Trends, one metric and goals are tabs inside My Coach. Reports, review, leaderboard and keywords are tabs inside Team Coaching. The sidebar stays a list of places, not a list of features.
Everything else on these screens is a stock card, table, badge, callout, segmented control or progress bar. The range bar was derived from the same 12 step roles: gray-4 track, green-3 band on a green-7 hairline, red, orange or green marker, gray-9 baseline tick.
Out of range is red, edge of range is orange, in range is green, team data is blue. Amber is left for the active nav pill, the one solid CTA per view, links and the AI voice. A score is never painted amber, so amber never reads as data.
Coach suggests 3m 30s when the ideal is 3m 00s, then raises the target after five held meetings. A target you can hit this week beats a correct target you ignore.
One sentence of justification, the moments that justify it, and an override that wins and is attributed. Without those three, nobody trusts a graded call twice.
| FR-1 | Knowing which speaker is you | |
| FR-2 | The fifteen metrics and their ranges | |
| FR-3 | The per meeting Coach card | |
| FR-4 | Moments, playback behind every metric | |
| FR-5 | Trends | |
| FR-6 | Recommendations | |
| FR-7 | Goals and the practice loop | |
| FR-8 | Weekly coaching digest | |
| FR-9 | Scorecards | Parked for this round |
| FR-10 | Manager coaching workflow | Parked for this round |
| FR-11 | Leaderboards and benchmarking | Parked for this round |
| FR-12 | Best practice clip library | |
| FR-13 | Keyword and topic tracking | Parked for this round |
| FR-14 | In person capture | |
| FR-15 | Administration | |
| FR-16 | Privacy | |
| CRM | Prospects, Discovery Coach, activities and notes | |
| L-1 to L-17 | Lisa, the digital twin assistant | |
| D-1 to D-11 | One-click AI draft and send | |
| US-6, US-7 | Role play, stage based practice | |
| US-10 | Sales guidelines behind the practice |
Team Coaching assumes a manager to direct report edge exists. The design shows six reports and states the rule in the blue callout. Wherever that edge comes from, the interface does not change.
Ranges are editable everywhere and Settings keeps the shipped default beside any edited value, so validating them against the corpus later costs nothing in the interface.
Scorecards, leaderboards, the clip library and keyword tracking are independently switchable modules, so a paid add on is a flag change rather than a redesign.
Filler lists and wpm bands are workspace editable in Settings. A per language set is the same screen scoped differently.
The interruption metric states its source and a confidence label, high, from timeline events rather than transcript overlap, which works whichever source ships.
Listed explicitly in the Settings exclusion list and marked pending, so the decision has a home in the interface either way.