← All Projects
B2B SaaS1 min read

Designing for Faster Collections in Autonomous Collections with AI

Redesigned the core collections workflow so AR analysts could act faster, combining AI insights, in-app calling, automated summaries, and a structured follow-up system.

40%
Fewer missed follow-ups
15%
Less manual effort
13%
Faster DSO
RoleProduct Designer (End-to-end)
PlatformFintech & AI, B2B SaaS
Team1 designer, 2 PMs, 4 engineers (2 FE, 2 BE)

Team Composition

Design1 Designer (me, end-to-end)
Product2 PMs
Engineering4 Engineers (2 FE, 2 BE)
AIFintechCollectionsWorkflow

TL;DR

Problem

AR analysts manage hundreds of accounts per cycle, but the workflow was fragmented across tools, leading to manual work, errors, and missed follow-ups that impacted DSO.

What I did

Redesigned the end-to-end collections workflow by integrating AI-assisted insights, in-app calling, automated summaries, and a post-call follow-up system.

Impact

Fewer missed follow-ups, less manual effort, faster DSO.

Designing for Faster Collections in Autonomous Collections with AI

Challenge

Discovery

Outcomes need structure. If they aren't captured in a consistent format, they get lost. An empathy map built from the discoveries showed analysts switching between tools constantly, losing context between prep, the call itself, and the follow-up.

Strategy

Decision 1: Unify the workflow in one place

Shipped

Why: Fragmentation created context switching and missed follow-ups.

  • Core dashboard now supports prep → call → log → next steps in one continuous flow

Result: Reduced manual effort and friction, both qualitatively and in time spent per account.

Decision 2: Bring calling inside the product

Shipped

Why: Calls are central, but the work after calls was the real bottleneck.

  • In-app calling with call context visible during the interaction

Result: Analysts no longer lost context switching between a separate dialer and the workflow tool.

Decision 3: Automate summaries to remove admin work

Shipped

Why: Notes and summaries were repetitive and inconsistent across analysts.

  • Automated call summaries that analysts can review and edit quickly

Result: Consistent, structured call records without the manual write-up.

Decision 4: Make follow-ups structured

Shipped

Why: Missed follow-ups directly impacted DSO, and ad-hoc reminders weren't reliable.

  • A structured post-call follow-up system tied to the worklist

Result: Fewer missed follow-ups and faster days-sales-outstanding across the collections operation.

Process & Visuals

Designing for Faster Collections in Autonomous Collections with AI: process visual 1
Designing for Faster Collections in Autonomous Collections with AI: process visual 2
Designing for Faster Collections in Autonomous Collections with AI: process visual 3
Designing for Faster Collections in Autonomous Collections with AI: process visual 4
Designing for Faster Collections in Autonomous Collections with AI: process visual 5
Designing for Faster Collections in Autonomous Collections with AI: process visual 6
Designing for Faster Collections in Autonomous Collections with AI: process visual 7

Results

Reflection

The real cost of this workflow wasn't any single broken step. It was the constant context-switching between tools that made every account take longer than it should have. Unifying prep, calling, and follow-up into one continuous flow mattered more than any individual feature inside it.

JARVIS