Brillianse CRM.

1. Lead arrivesMeta and Google ads, branded landing pages, WhatsApp
2. AI calls in 5 secondsVoice agent qualifies the lead and fills the form
3. RM on a coached callBridged call, live suggestions, sentiment
4. PaymentLead becomes an applicant, then pays
5. Student in BrillianseRecord moves to the admin panel, no retyping

The problem

Admissions at an Indian junior college run on lead volume, and a lead goes cold fast: a parent who does not hear back calls the next college. Three things decide whether an ad lead becomes an admission: how fast someone calls, how good that call is, and whether the family's details survive the hand-off into the academic system.

Each institution also needed its own face on all of it: its own logo, colours and landing pages, and its own phone numbers, so a parent deals with the college, not a vendor.

What I built

Speed to lead: an AI call within five seconds

Every channel lands in one queue: Meta lead ads, Google ads, the branded landing pages and WhatsApp. Within five seconds of a new lead, the AI voice agent calls it, takes the parent through a short script, and fills the enquiry form from the conversation. It scores the call and writes it up, so the RM who picks the lead up next starts with the details already captured.

Calls the way an RM actually works

• Click to call from the laptop. An RM calls a lead straight from the CRM in the browser, over the institution's own number.
• Bridged calls. The agent calls the RM first, then dials the parent and joins the two, and it listens to both sides of the conversation.
• Live coaching. While the call runs, the RM sees suggestions on what to say next, drawn from the live transcript, with a running read of the parent's sentiment.
• On their own handset. A companion app lets counsellors take coached calls on their phones, with the audio and a per-speaker transcript kept for review.

Brand Factory: each institution's look, on demand

Brand Factory holds every institution's assets (logos, colours, copy) behind a prompt-driven editor in the style of Lovable: describe a change and it is made against that brand. From the same assets it generates landing pages and shareable links on the fly, and AI-drafted ad creatives.

Ads in, leads back

The CRM connects to each institution's Meta and Google ad accounts through OAuth, using a Facebook app configured on our side. Creatives and landing links made in the CRM go out as ads, and the leads those ads produce come straight back into the CRM queue, where the five-second call picks them up.

From lead to student

A record moves from lead to applicant to paid. When the details are complete and the payment clears, it becomes a student in the Brillianse admin panel: a status change on the same data, not an export and a retype. Campaigns go out over SMS and the WhatsApp Cloud API with consent gates, unsubscribe and delivery receipts, and repeat enquiries, both parents and siblings are linked into one family.

Telephony

Twilio is the main carrier, and Plivo runs alongside it, since this was an Indian client. I set up the SIP trunks and bought numbers for each institution, so every call, from the AI agent or an RM, goes out on that institution's own line.

Trade-offs

• Building the CRM rather than buying one. An off-the-shelf CRM would have been running months earlier. It would also have sat outside the platform's auth, data model and reporting, and an admissions conversation would never have joined up with the student record it becomes.
• The AI calls first, the RM follows. A machine call in five seconds beats a human call in an hour, so the agent makes the first contact and captures the details, and the RM spends their time on the conversations that need a person.
• Coaching on screen, not in the ear. Suggestions appear on the RM's screen to use or ignore, so the conversation stays the RM's own.
• Two carriers. Twilio stays the main carrier for its tooling, and Plivo adds Indian numbers for an Indian client, at the cost of setting up and running trunks on both.