Case study 04 · Marketing automation
WhatsApp automation for an intern-agent programme
A bilingual WhatsApp sequence on Interakt designed to recruit each internship cohort — and the CRM model that tracks graduates as referral agents, from commission to payout.
Tools
mego_intern_final_hours · reusable draft
Last call, {{1}}. Registration and the entry exam for cohort {{2}} both close at the end of {{3}}.
After that, the next opportunity is a future cohort. Registering is free and takes a few minutes.
If something is holding you back, just reply here and someone from our team will help.
Overview
MEGO Forex runs an internship programme. Students register, complete KYC and sit an entry exam, train as agents, and can then refer customers — earning commission on the business they bring in.
It’s recruitment, onboarding and a sales channel in one. WhatsApp is where the conversation happens; the CRM is where it’s recorded, attributed and paid.
The system
From first message to payout
- 01 Reach WhatsApp, social ads and referrals.
- 02 Five-message sequence Interakt English and Hindi, approved by Meta.
- 03 Register Free registration per cohort.
- 04 KYC & entry exam
- 05 Trainee → graduate Tracked as agent stages in the CRM.
- 06 Agent refers customers Zoho CRM Each lead carries its referring agent.
- 07 Commission → payout Pending, approved, paid or disputed — per agent.
The message sequence
Five approved WhatsApp marketing templates, each in English and Hindi:
- First contact — the opening message.
- Programme info — what the internship involves.
- How to join — how to register and sit the entry exam.
- Deadline reminder — registration is closing.
- Final hours — last call, with an open invitation to reply to a person.
Every template carries an opt-out line, and each message has one job.
Built to be reused
Reusable version I drafted · cohort and dates as variables
mego_intern_deadline_reminder · English
Quick reminder, {{1}} — registration for cohort {{2}} closes on {{3}}.
The entry exam window closes the same day, so if you want this cohort it is worth registering today rather than leaving it. Registration is free and takes a few minutes.
Results are published on {{4}} and the internship runs {{5}}.
Variables
- {{1}}
- Recipient’s first name
- {{2}}
- Cohort ID
- {{3}}
- Registration and entry-exam deadline
- {{4}}
- Results date
- {{5}}
- Internship dates
Making it reusable
The first versions of the templates had the cohort ID and every date written directly into the copy. That works once. For the next cohort, every template would need rewriting and re-approval by Meta.
I reworked the copy so the cohort, deadline, results date and internship dates become variables filled in at send time — so each template could be approved once and reused for every cohort. I also flagged the one thing that can’t be a variable: Meta doesn’t allow variables in a button’s label, so any date in button text has to go and the call to action has to be written date-free.
Consent and the human handoff
- Consent is recorded, not assumed. The CRM stores WhatsApp, SMS and email consent separately, with the source, the version of the privacy notice shown and a timestamp.
- Each lead records who handled the conversation — the AI bot, a person, or escalated — so nobody falls through the gap between automation and the team.
- A person is one reply away. The final message ends: “If something is holding you back, just reply here and someone from our team will help.”
The agent data model
Custom Zoho CRM modules I built
- 01 Cohort Dates, capacity, track (stipend or certification) and status.
- 02 Agent KYC status, referral code, and stage: registered → KYC verified → exam submitted → trainee → graduated → agent.
- 03 Agent commission Linked to the agent and the deal; pending, approved, paid or disputed.
- 04 Agent payout Gross, TDS (tax deducted at source), deductions and net, with payment status.
Outcome
The templates are approved in English and Hindi, a reusable version is ready for future cohorts, and the CRM has consent and handling fields on every lead — plus a data model that links each commission to its agent, deal and payout.
Capabilities demonstrated
WhatsApp Business template design · working within Meta’s platform rules · bilingual messaging · consent-first automation · bot-to-human handoff · data modelling for a commission-based channel.