Hybridge Implants

Consultation QA

Sign in with your @hybridgeimplants.com or @elmwooddental.com account.

Hybridge

Consultation QA

+ New Submission
All Runs
Analytics
Settings

Score a recording

Recent meetings


Or score a specific URL

Paste a Zoom recording share link, the meeting's Play URL, or the numeric Meeting ID. The pipeline resolves it, fetches the transcript, runs Gemini, and produces the report.

Recent runs

When Doctor Patient TC Location Score Status Trigger Logs
No runs yet. Auto-triggered ones appear here as Zoom recordings finish; submit one manually with the +New Submission tab.

Doctor performance

Aggregates over completed runs.

→
Locations
Doctors

Insight reports

Constructive, comparative coaching across doctors — saved to the spreadsheet + PDF.
GeneratedDoctorsRangeConsultsStatus
No insight reports yet. Click “📊 Analyse…” to generate one.
No completed runs in this window. Submit a few from + New Submission.

Summary

Category score heatmap

Per-criterion averages. Color-coded against the 7-point bands you'd expect: ≥9 great, 7–8.9 solid, ≤6.9 needs work.

Category scores — radar

Toggle doctors and locations above; the Window selector sets the time range for both charts.

No doctors selected. Turn some on above.

Trend — total score %

Each doctor's average total as a percentage. Use the chips above to filter who's shown.

No doctors selected. Turn some on above.

Case outcomes from Monday.com

Coverage and outcomes

Counts of consults in the current window. Status columns cover every consult with a Monday card (confident and low confidence together). Click a row to see the patients behind it.

Cohort view

Same consults as percentages. Match quality is a share of graded consults; each status block is a share of its own tier, so confident and low confidence matches are never mixed. Click a row for details.

Match workbench

Confirm or fix every consult the matcher was not sure about. Validated links update the master sheet name and survive weekly re-imports.

Show
Select a consult on the left to review its match or find the right Monday patient.

Patients & case outcomes

Every graded consult with its patient id, scores, and Monday.com case status. Click a status chip to filter.

No consults match the current filters.

Won vs lost — score profile

For each location and doctor: the average score on consults where the patient accepted treatment, and next to it a colored pill showing how much higher ▲ or lower ▼ that is than their consults where the patient decided against it. Big green pills mean that step is where their winning consults pull ahead. The two money columns come from Monday: avg proposed is the average Tx Value across every decided consult that carries one, won or lost, and avg accepted is the average on the ones that closed. A red pill there means the bigger cases are the ones getting away.

Which scores travel with a yes

The link strength pill says how strongly a step's score moves together with the patient saying yes: green + means higher scores went with more acceptances, red − means the opposite, gray means no clear link. Then the practical version: the acceptance rate when that step scored well (6+) vs poorly (5 or less). Only decided patients count. A link is a pattern, not proof of cause, and pills on few cases (small n) are just a hint.

Conversion charts

Accepted the treatment Decided against Still deciding / no card

Where winning consultations score higher

Each pair of bars is one step of the consultation. The green bar is the average score when the patient went on to accept treatment; the red bar is the average when they decided against it. The wider the gap, the more that step separates a won consult from a lost one.

Do better-scored consults actually convert?

Consults grouped by their total score (out of 70). The bar shows what share of the decided patients in that band accepted. If coaching works, the right side of this chart should be taller than the left.

How every consult ended, by total score

The same score bands, showing every consult: green went on to accept, red decided against, gray is still deciding or has no Monday card yet. Tall gray means follow-up opportunity, not failure.

Month by month: quality vs conversions

Green line = the share of decided patients who accepted each month. Purple line = the average consultation score that month. When the lines move together it suggests better consults are paying off.

Acceptance rate by doctor

Of each doctor's patients who made a decision, the share that accepted. Hover a bar to see the counts behind it — a high rate on very few cases means less than it looks.

Outcome mix by location

All consults per office, stacked by how they ended. Compares the shape of each location's funnel at a glance.

Better consults, faster yes?

One dot per accepted case: its consultation score across, and how many days after this consult the patient accepted, up. Dots drifting down-right would mean stronger consults close sooner. Days are counted from the scored consult (not from any older first visit on the Monday card), and cases where the yes came before this call (follow-up visits after an earlier acceptance) are left out.

Consultation score vs treatment value

One dot per accepted case: consultation score across, dollar value of the accepted treatment up. Shows whether bigger cases get (or need) better consultations.

Bigger cases, longer decisions?

Accepted cases grouped by treatment value. Bars show how many days the average patient in each price band took to say yes after their consult; the line shows how many cases each band holds. Tells you whether the expensive cases need a different follow-up rhythm.

Zoom accounts

Each location or organization with its own Zoom login connects here. Connecting an account lets the pipeline receive its recordings.

Mailboxes and coordinators

Every Zoom mailbox that records consultations needs a row here. The row decides the location label, who the coordinator is, and who receives the report and clinical note.

Scoring and note prompts

The AI prompts, model choice, and creativity settings live in their own editor.

Run detail

Scoring Settings

Edit how the AI evaluates consultations. Changes go live within ~5 minutes — no redeploy needed. Anyone with access can maintain this; just write plain English instructions.

i

🧠 Which AI model to use

Higher quality models cost more and take longer. Gemini 2.5 Pro is the default.

🎭 How the AI behaves

A short description of the AI's persona and what it cares about while scoring. Click ✏️ to edit.


    

📋 Rubric — general instructions

What gets graded overall, plus any cross-criterion rules. Each individual criterion gets its own editor below — keep this section focused on the big picture (scoring style, JSON output reminders, etc.). Meeting details and the full transcript get auto-attached — don't mention them here.


    

🎯 Per-criterion instructions

Edit any single criterion below without touching the others. The colored border on the left of each card matches the same color used in the consultation report so non-technical reviewers can pattern-match easily. Hover the ✏️ icon on any card to open its edit dialog.

📝 Final notes & output format

Anything that should come AFTER all seven criterion blocks — output format reminders, JSON schema guidance, etc.


    
⚙ Advanced — generation knobs

🩺 Clinical note generation

After scoring, a second AI pass turns the transcript into a clinical chart note and saves it as a Google Doc in the patient's Drive folder. These settings control that note — they're independent of the scoring rules above.

🧠 Which AI model writes the note

Defaults to Gemini 2.5 Pro. Drop to a Flash model if you want it cheaper/faster.

📝 Note-writing instructions

The full instructions the AI follows to write the chart note (structure, tone, terminology rules). The meeting details and transcript get auto-attached — don't paste them here. Click ✏️ to edit.


      
⚙ Advanced — generation knobs

Analyse Consultation consistency insights

Pick doctors/combos and a date range. We generate per-doctor strengths, weaknesses, and a constructive comparison, save it to the spreadsheet, and produce a downloadable PDF.

→
Doctors / combos
Locations

All locations are included unless you untick some.

Starting…

0%

You can close this window — the analysis keeps running and shows up in the Insight reports table.

Insight report

Email this report

Sending is optional — the report is already saved. Pick recipients:

Outcomes

Confirm Is this the right match?

Edit Edit prompt

Edit the prompt text below. When you're happy with it, hit Review changes to see exactly what's different from the live version before publishing. Want the AI to clean it up first? Hit the ✨ button in the textarea's bottom-right.

Review the differences from the live version. Removed lines are red, added lines are green. Hit Confirm & publish to make this live, or go back to keep editing.

ⓘ