Routed
Conversations reaching Mia
466
92 at 07-18 audit → +407%
Info
Reached the program-info stage
311
67% of routed
Commit
Committed to show up live
163
19 at audit · evaluator-confirmed
Leak
Warm leads dropped (unrouted)
22
missing ManyChat keywords
How to read this (plain-language glossary)
Routed
Conversations that actually reached Mia. Ninjo calls this "Agent Conversations". The thousands of others are broadcast sends with no reply.
Program info
Lead got as far as the program details / link stage in the chat. The closest thing to a CTR stage.
Committed
Lead said they'll show up live to the webclass. This is our main conversion signal for a show-up funnel.
Conversion rate
A stage count ÷ the routed base. We divide by routed, not by all the broadcast sends, so the % reflects real conversations.
Leak
Leads who replied on WhatsApp but never reached Mia — mostly because they typed a keyword (e.g. "cowork") we haven't wired up yet.
Funnel (466 routed conversations · Ninjo funnel steps · % of top · % vs previous)
Contactedrouted & engaged
466
100%
Prequalificationshared goal / pain / experience
442
95% · 95%
Nurturingresource shared / received
441
95% · 100%
Program Infoprogram details & interest
311
67% · 71%
Booking Sharedwebclass link, phone or email captured
122
26% · 39%
Conversion rates (base = the 466 conversations that reached the agent, not the broadcast sends)
Reached program info · "CTR"
67%
311 / 466 routed
reached the program details / link stage
headline
Conversion rate · committed
35%
163 / 466 routed
committed to show up to the webclass live
Info → commit
52%
163 / 311 reached info
of those who reached program info, how many committed
Confirmed outcomes (evaluator-filled boolean properties, label-keyed)
Booking link sent
111
confirmed sends
Show-up committed
163
terminal conversion
WhatsApp opt-in
33
reminder-eligible
Not interested
4
agent stops
Daily — conversations routed to the agent (the number under each bar is unrouted that day)
Routed to agentunr = unrouted (mostly outbound-only broadcast sends)
Who is showing up & what they want (evaluator segments across routed conversations)
| Buyer persona | Leads |
|---|
| non-technical aspiring AI consultant | 110 |
| aspiring AI consultant | 32 |
| busy professional | 7 |
| non-technical beginner | 7 |
| non-technical mum | 2 |
| mum of two | 2 |
| Goal shared | Leads |
|---|
| get paid helping businesses with AI | 71 |
| get paid to help businesses with AI | 67 |
| get paid $5K–$10K+/mo helping businesses with AI | 10 |
| learn how to get paid helping businesses with AI | 3 |
| grow | 3 |
| get a client that wants to pay for the service | 1 |
| Top pain point | Leads |
|---|
| finding clients | 2 |
| getting a client that wants to pay for the service | 1 |
| how to get clients and ensure they pay | 1 |
| SMB AI adoption gap | 1 |
| AI is fast evolving | 1 |
| connecting with potential clients without sounding like others pitching AI solutions | 1 |
| Objection raised | Leads |
|---|
| finding clients | 2 |
| needs another day besides Friday | 1 |
| noisy market / standing out from other AI pitches | 1 |
| price curiosity | 1 |
| too expensive | 1 |
| cannot attend live | 1 |
Routing leak — 152 unrouted replies (leads who replied but never reached Mia · recovery list)
| First inbound | Count | Status | What it is / action |
|---|
| cowork | 15 | Warm · fixable | ManyChat keyword, not a Ninjo trigger → sync it |
| consultant | 4 | Warm · fixable | ManyChat keyword, not a Ninjo trigger → sync it |
| playbook | 2 | Warm · fixable | ManyChat keyword, not a Ninjo trigger → sync it |
| launch | 1 | Warm · fixable | ManyChat keyword, not a Ninjo trigger → sync it |
| "20k in 12 weeks👏" · "ah i see" · "bonjour je suis à afrique" | 51 | Warm · free-text | real intent, no keyword — hard to catch |
| yes · aic · grow · live … | 24 | Now routing | registered triggers; residual mostly pre-fix |
| stop · delete my number | 10 | Ignored ✓ | opt-outs, correctly left alone |
| business autoresponders / links | 45 | Noise | other companies' WhatsApp bots, not leads |
What this says
✓
The setter works where it engages. 466 routed → 311 reached program info → 163 committed to show up live, up from 92 routed / 19 committed at the audit. Voice is clean, objections deflect to Friday, no leaks.
!
The keyword sync isn't finished. The yes fix helped, but cowork / consultant / playbook / launch are the real dropped bucket (~22 warm leads) — ManyChat keywords that need syncing as Ninjo triggers. Same fix, wider net.
→
The show-up engine is the next lever. "Can't attend live / wants replay" is the one repeating objection cluster, and 163 committed leads have no reminder cadence. WF-2 (WhatsApp 3d / 1d / 3h / 15min) converts commitment into attendance.
→
Operators are flying blind on wins. 163 leads committed to show up, but there's no notification when it happens — the operator would have to watch DMs. One upsert_custom_notification on show_up_committed closes that.