Skip to content

Internal tool · ads to WhatsApp to closes · Lahore

CoRelations growth dashboard

Ads, WhatsApp conversations and closes in one funnel, with the AI agent's lead quality measured, not guessed.

CoRelations · Growth

Live · hourly
Ad spend
PKR 0
last 30 days
Conversations
0
WhatsApp chats started
Real workspace leads
0
qualified by Sarah
Cost / booked tour
PKR 0
the number that matters

Funnel · impression to close

  1. People reached0
  2. Link clicks0
  3. Conversations0
  4. Real chats (3+ msgs)0
  5. Workspace leads0
  6. Closes0

A chat is not a lead. Sarah reads every one and sorts them before they count.

Conversations per dayWorkspace leads per day

Who is actually messaging

  • Here for workspacethe ones we want
  • Asking for a jobfiltered out
  • Just browsingfollow-up later
  • Wrong number / spamignored

Sample data · real figures stay private to the client

Problem

Click-to-WhatsApp ads report conversations, but a conversation can be a job seeker or a wrong number. The team needed to know which ads and audiences brought people who actually wanted a desk.

What we built

A dashboard that joins Meta ad data with the conversations Sarah (the AI front desk) handled and the team's closes log. It classifies who is really messaging, runs a proper audience test, shows the best hours to advertise, and tracks the website in a separate tab.

Result

Spend is judged on cost per real workspace lead and cost per booked tour, not on cost per chat. Audience winners are decided with a statistical test, not a hunch.

The starting point

A coworking space sells through WhatsApp. Meta counts every chat the same, so the cheapest ads looked best even when they mostly attracted people asking for jobs.

What we built

An hourly pipeline that pulls ad data, reads which conversations Sarah qualified, and lines them up with paying members from the closes log. Every section explains what it is measuring in plain language.

How it is used

The team decides budgets on cost per real lead and cost per booked tour, and checks the hour-by-hour view before scheduling ads. The dashboard shown here uses sample data; the real figures stay private.

Next case studyTradeX Student Portal

Let's build
yours

Reply within one working day.

Rather see numbers first? Get an instant estimate