Cases No 01 2026

Autoflow A debt-collection platform for a Brazilian credit company.

Built for a consumer-credit company in Brazil. The team had been managing collections across spreadsheets, WhatsApp, and a legacy CRM. We consolidated the workflow into one interface and used AI to reduce repetitive operational work.

portfolio today 142 aging 38 at risk 12 0–7d 8–30d 31–60d 60d+ 01 Dashboard portfolio · prioritised contract · 0428‑3 issued due +7d +30 AI · contract read next · settlement offer 02 Contract live data · AI read inbox · triage PAY NEG PAY N/R 03 Triage intent · reply · next
Fig. 01 The three screens a collector uses day to day. The dashboard shows where to start, the contract view holds the debtor context, and the triage view sorts inbound messages by what each one needs.

Numbers What changed

  • i ~50% of a collector's day used to be operational work. Most of that time is now available for customer conversations.
  • ii 2 days minutes Invoice issuance used to take the team two days each cycle. It now runs automatically.
  • iii ~3x Target cases per collector with AI handling repetitive work. The system is tracking toward that target.

No 01

The problem

When we started, the team's day was approximately half negotiation and half operations. The operational half included issuing boletos, copying contract numbers between systems, reading inbound messages to identify next steps, and updating statuses afterwards. Many collection teams operate this way.

The brief was to give them the operational half back.

No 02

What the AI does

AI is embedded in four workflow steps, rather than presented as a standalone chatbot.

  1. i

    Triage

    Inbound messages on WhatsApp and SMS are tagged by intent: payment, negotiation, no response, question, complaint. The triage view groups them by tag so collectors can process high volumes without constant context switching.

  2. ii

    Reply

    When a collector opens a conversation, a suggested response is drafted from the relevant contract context. Instalment numbers and payment details are pulled at draft time.

  3. iii

    Contract read

    Each contract gets a short summary: what happened, what is overdue, and the recommended next action. This was the first capability requested by the team managers.

  4. iv

    Flows

    A no-code builder for SMS, WhatsApp, and voice campaigns. Flows can be triggered by time, calendar, or contract events such as invoice issuance, payment receipt, or days overdue. Managers can run campaigns without engineering support for each change.

inbound WhatsApp · SMS AI · classify intent → lane ground live contract AI · reply suggested · sent instalments · settlement · history · status
Fig. 02 What happens when a message comes in. The system classifies the intent, retrieves the relevant contract context, and drafts a grounded reply. The collector reviews, edits if needed, and sends.

No 03

Where it's going

We planned the product in three stages. The first stage is in production today. The next two define the roadmap for the coming year.

  1. 01 Shipping

    Copilot

    Where we are today. AI drafts the reply; the collector reviews it, edits if needed, and sends. A human remains in every loop.

  2. 02 In progress

    Assisted

    AI handles repeatable patterns such as reminders and simple promise-to-pay confirmations, then asks for approval when negotiation or discount rules are involved.

  3. 03 Planned

    Autopilot

    Most routine cases can close without manual review. The exceptions queue stays small enough for the team to handle carefully.

No 04

The stack

Surfaces
Dashboard, contract view, conversation triage, flow builder, team monitor.
AI
Intent classifier, reply drafter, contract summariser, decision branches inside flows, AI voice for outbound calls.
Channels
WhatsApp, SMS, and AI voice, all running through the same flow engine.
Data
The platform talks to the company's credit system in real time. Nothing the AI sees is more than a few seconds stale.
Live today
Two AI agents in production. Four automated flows running.