Our pilot showed contact rates of 33% on WhatsApp, 8% by email, and 19% by SMS at day seven of delinquency. The Q4 2025 test covered approximately 2,400 accounts across three lender portfolios. Yet most collections software sold in Brazil still defaults to email-first.
This post explains why that default persists, what it costs, and what a working Brazilian channel strategy requires technically.
Why Email-First Collections Fails in Brazil
Email-first reflects US and European assumptions built into many collections platforms. There, email offers inbox access, mobile notifications, reasonable financial open rates, and the documentation those tools were designed to support.
In Brazil, that premise weakens. Borrowers use WhatsApp throughout the day, while lender emails often land in filtered tabs or go low-priority. WhatsApp feels like conversation. Email feels like a notice.
Our pilot and wider market observations point to WhatsApp collections opens running approximately four times higher than email. More useful is payment response: WhatsApp-first sequences produced a first-payment response rate 2.8 times the lenders' prior email-only baseline over an equivalent 30-day window.
WhatsApp Mechanics That Matter for Collections
WhatsApp alone is not a strategy. Message structure, approval, and sequencing determine whether it beats email for a portfolio.
WhatsApp Business API requires approved templates for initial outbound contact under Meta's business messaging policies. Templates should be concise, use one clear call to action, and avoid mass-broadcast language. Borrowers distinguish a personal message from one sent to 5,000 accounts, and response rates reflect it.
Short form matters. A 400-word dunning notice suited to email fails on WhatsApp. Use 80-120 words and one next step: amount, overdue period, and one path, such as a PIX link or a reply to discuss options.
Response handling is often underestimated. A borrower who replies and hears nothing for 24 hours may have a worse experience than one who received no message. Use an automated flow or a human queue with a defined SLA. Outbound WhatsApp without inbound handling can raise contact rates while leaving resolution rates behind.
LGPD applies to WhatsApp channel data. Preference signals, engagement history, and behavior from these interactions are personal data under Lei No. 13.709/2018. Document the processing basis for routing, typically legitimate interest under Art. 7, IX, with a balancing test.
Routing Signals That Predict Channel Preference
Sending every account to WhatsApp beats email-first, but remains batch thinking. The better question is which channel gives this borrower the highest contact probability now, based on engagement history.
The BLU365 routing model uses four signal types.
First, prior engagement. A borrower who answered the lender's WhatsApp service channel is a strong WhatsApp signal. Lenders should pass WhatsApp service data from core banking or CRM systems to collections routing.
Second, prior-cycle clicks and interactions. A borrower who repeatedly clicked SMS payment links may respond to SMS even with low WhatsApp engagement. Preference often persists across delinquency cycles.
Third, timing. Some borrowers engage during the morning commute, 7-8 AM; others in the evening, 7-9 PM. A borrower active on WhatsApp at 8 PM but unresponsive to midday SMS should receive an evening WhatsApp message, not the 10 AM daily batch.
Fourth, balance tier. Give higher-balance accounts priority when WhatsApp template capacity is limited. Lower-balance bulk segments can use email or SMS without significant recovery loss.
Years of history are not required. Two or three attempts in the current cycle can update preference predictions. Each touchpoint improves the sequence.
Where Email Prevails
Email is not obsolete in Brazilian collections. It handles some jobs better than WhatsApp, so a sound strategy uses both.
Email suits documentation and formal disclosure. Payment agreements, statements, and cure notices need a durable, auditable record. Some compliance programs require overdue notices by email even without an open, supporting later escalation or credit bureau reporting under CDC Art. 43.
Email also suits high-volume, low-balance accounts. A segment of 1,200 accounts below R$300 does not merit WhatsApp capacity reserved for higher-value recoveries. An email sequence with a PIX link can perform adequately at lower cost per contact.
In practice: WhatsApp first for medium-to-high balances with supporting engagement data; SMS where WhatsApp history is absent but SMS response exists; email for formal documents and low-priority volume. Each channel serves a different portfolio need.
Building the Routing Layer: Technical Needs
A lender adding channel intelligence must choose build or buy. Building needs an API layer for WhatsApp Business API, SMS, and email, an account-level event store for delivery, opens, and replies, and a preference model that updates channel scores.
For a team with dedicated data engineering, this is a three-to-four month project. For an engineering team focused on origination, credit decisions, and core banking, it can distract from the product roadmap.
BLU365 connects overdue portfolios by API or CSV. The platform runs routing, message generation, response handling, and payment confirmation. Its dashboard shows recovery rate, contact rate by channel, days-to-resolution, and channel efficiency, while the lender's engineers maintain none of the routing mechanics.
To test the gap, run a 30-day test on the next overdue cohort. Send half through the current sequence and half through a WhatsApp-first sequence at equivalent timing. Contact rates should separate within two weeks, with recovery conversion clearer by the end.