Tool · Collections
Collect more, chase less.
Set a collection strategy for each risk level (which channels, how often, how firmly) and let AI run it: text messages, calls and notifications at the pace you define, while your team focuses on the cases that need a person.
Why it matters
Everyone gets chased the same way.
A customer one day late and one ninety days late end up on the same call list.
Payments are missed that never needed to be.
Nobody reminded the customer before the due date.
Your team can't call everyone.
So the calls that matter most wait behind the ones that didn't need a person.
What it does
The loan book by days late
Every loan grouped by how late it is, live, with the amount at risk in each group.
Loan book by days late · Consumer loansLive
Loan book 3,820 loans · $17.6BLate 408 loans · $2.4BUpdated as each payment arrives
Risk rulesYour criteria
Days late1–30 days → Medium · 31+ → High
Amount dueOver $5,000,000 → one level up
Payment historyTwo or more late payments → one level up
Your own criteriaProduct, region, segment…
Low riskRun by AI
2,184 loans
Step up to Medium after a missed payment.
Medium riskRun by AI
296 loans
Step up to High after 10 days with no payment.
High riskRun by AI
112 loans
Cases you choose go to your team's work list.
For your teamThe cases you assign to people, ranked by likelihood of payment. Everything else runs in the strategies.
Payment promiseCamilo P. promised $275,000 for March 14 · recorded by D. Rojas, visible to the whole team
Reminders · LN-38812 · Camilo P.Paid
3 days beforeText message that the payment is coming upSent
Due dateNotification with a link to paySent
1 day latePayment missed. Loan moves to 1–30 daysLate
2 days lateText message with a link to paySent
2 days latePaid from the message. Back to currentPaid
Customers pay through your portal or a ready-made payment integration.
How it works
AI in this tool
AI runs your strategy. AI agents make the calls and send the messages and notifications for each risk level, at the pace you set.
Who to call first. AI ranks late loans by how likely each customer is to pay.
Early warning. AI flags loans that are starting to slip before they're late, so you can act early.
Payments from the conversation. On WhatsApp, the AI assistant can take the payment right in the chat.
Use it with Cowally, or on its own.
Use it from the back office, connected to every other part.
Your developers plug it into your own product.
See API By MCPYour people and AI agents reach it, with roles and rules.
See MCPQuestions about collections
Yes. You define the levels, the rules that place each loan in a level, and each level's channels and pace.
Text messages, phone calls, messages and notifications, in the combination and at the pace you set.
Yes. AI agents make voice calls and send the messages for each strategy; your team handles the cases you assign to people.
Yes, through your portal or a ready-made payment integration.
Yes. Your systems send the loan book through the API, and collection results come back.
Yes, with the promised date and amount, so everyone sees what was agreed.
See your loan book, with a strategy for every risk level
In one session we'll set up strategies for three risk levels with a loan book like yours, and show AI running them.