← Agent catalogue

AI Agent Profile · LendingIQ · Bengaluru

Field Collections Coordinator AI

Function: Field Collections SupervisorInvoked via: collections work-list — daily field planning batchRuntime: AWS Bedrock · ap-south-1Model: Claude Sonnet 4Context window: 200K tokens

DivisionCollections

Resume

What this agent does

The Field Collections Coordinator AI optimises LendingIQ's field collections programme — building the daily route and account assignment plan for each field agent, matching agent profiles to account types, logging visit outcomes as field agents report back, and identifying accounts where repeated field visits have produced no results and a different strategy is needed. It turns a manual planning process into an optimised, data-driven operation that maximises field agent productivity while respecting RBI contact hour and conduct requirements.

Primary functions

Route Optimisation

Daily — before field agent deployment

INVOKED WHEN: daily field planning batch runs, typically by 7 AM before agents deploy

  • Reads the day's field-eligible account list — accounts referred by the Mid or Hard Bucket Agent AI as unreachable by phone or requiring a physical presence — and groups them by geographic cluster, building the most time-efficient route for each agent that covers the maximum accounts within their zone, respecting the 8 AM–7 PM contact hours constraint.
  • Applies priority weighting within each route: higher DPD and higher outstanding accounts rank first, ensuring that if an agent cannot complete all assigned visits, the highest-priority accounts are visited first.
  • Checks CBS payment status immediately before including each account in the route — an account that paid overnight is removed automatically, preventing an unnecessary visit to a borrower who has already paid.
Output: Daily route plan per field agent — accounts in visit sequence with address, priority, account brief, estimated visit time, and CBS payment status confirmation.

Agent Assignment

Daily — matched to account type and zone

INVOKED WHEN: route planning is complete and agents need to be matched to account clusters

  • Matches each account cluster to the most appropriate available field agent based on geographic zone, language capability, account type familiarity, and prior visit history — where an agent has previously visited and built rapport, continuity improves outcomes.
  • Tracks agent capacity and does not over-assign — an over-assigned agent rushes visits, increasing the risk of conduct issues that arise from time pressure.
  • Blocks assignment to an account where the prior visit log records a borrower complaint about that specific agent — until the human field collections manager has reviewed and cleared the complaint.
Output: Agent assignment — daily account list per agent with visit sequence, account briefs, capacity utilisation. Complaint-flag blocks applied where applicable.

Visit Outcome Logging

Real-time — as field agents report back

INVOKED WHEN: field agent submits a visit outcome report via the mobile app

  • Reads the field agent's visit report — outcome code (met borrower / absent / PTP captured / payment collected / referral required), visit timestamp, and notes — and logs the outcome in the CRM against the account record.
  • For PTP captures: creates a PTP record with committed date, amount, and agent name, routing a WhatsApp reminder to the borrower on the committed date.
  • Flags accounts where three or more field visits produced no meaningful outcome for Collections Head AI strategy reassessment.
Output: CRM visit log with outcome code, timestamp, agent notes. PTP record where captured. Three-visit no-result flag to Collections Head AI.

Hard guardrails

Will notScript or instruct field agents on what to say to borrowers. This agent assigns routes; conduct is the human manager's responsibility.
Will notAssign field visits outside 8:00 AM to 7:00 PM. Visit windows are hardcoded.
Will notContact borrowers directly. All borrower contact goes through field agents or the bucket caller agents.

Known limitations

Route optimisation quality depends on address geocoding accuracy.Borrower addresses in Indian lending portfolios — particularly in rural areas — are often incomplete or ungeocodable. Build an address quality flag into the LOS at onboarding so addresses are verified before the borrower becomes a collections account.
The agent cannot verify that field visits actually occurred.GPS timestamping of visit reports — confirming the agent's device was at the account address at time of logging — reduces but does not eliminate false visit records. Build mandatory GPS confirmation into the field agent mobile app.
Agent Profile · Field Collections Coordinator AI · LendingIQ · BengaluruLast updated April 2026 · For internal use

Important Reads

Learn more about how to deploy Field Collections Coordinator AI to your lending workflow.