Features built for settlement control.
Matching, fee review, remittance tracking, and shared follow-up for Indian marketplace operations.
Multi-format ingestion
Drop in five spreadsheets, two PDFs, and a phone photo of a receipt. ReconPe unifies them with AI into one canonical dataset you sign off on before reconciliation.
Each source keeps its own schema — order_id, txn_ref, ORDER_NUMBER. An AI pass collapses them into a single canonical set of fields, shows you the full mapping for review, previews 10 sample rows, then materialises one clean file for matching. The reconciliation engine never sees the messy side.
Document Intelligence — PDFs, images, marketplace settlements
Upload a bank statement PDF, an Amazon settlement report, a photographed invoice, or a scanned receipt. Document Intelligence extracts structured tabular data on upload — AWS Textract for structured documents, Vision AI for handwritten and multilingual content. Included on every plan with progressive allowances.
A cascading pipeline routes structured documents (bank statements, invoices, settlements, receipts) to AWS Textract for deterministic, page-billed extraction; multilingual or handwritten content falls through to Claude / GPT-4o vision automatically. Per-page metering scales with the plan: Free 1 document (taste of the cascade), Starter 40, Growth 150, Pro 1,500 with overage. Confidence heatmap on every extracted row, and full audit trail at sign-off. Extracted rows slot into accumulation and reconciliation as if they had been CSV all along.
Cross-run memory
ReconPe remembers every exception by counterparty-amount-direction fingerprint. When this month's run turns up last month's open item, you close both with one click.
After three confirmed settlements for a counterparty, a pattern is tagged automatically — "late settler, avg lag 5 days" — and shown as an amber banner on future exceptions for that entity. Rejected candidate matches feed the next run's decisions, per organisation.
Guided investigation agent
Ask the agent to investigate any reconciliation and watch it think — a structured pipeline narrates each step via live Server-Sent Events, grounded in your data, not inference.
The agent runs a fixed investigation plan — match-rate distribution, exception type mix, risk signals, counterparty history — and streams its reasoning as it goes, so teams can answer "why did this run throw 40% exceptions?" without hunting through dashboards.
Marketplace auto-detection
Upload supported settlement exports and let ReconPe identify Amazon India, Flipkart, Meesho, Razorpay, Cashfree, PhonePe, or PayU formats automatically.
ReconPe recognizes the field structures and naming patterns of major Indian marketplace and payment reports so teams can skip manual mapping.
Commission variance review
Compare settlement lines against expected fee logic and flag likely commission mismatch with clear line-level context.
Category-wise or channel-specific fee rules can be checked against actual settlement outcomes so teams can spot likely overcharges sooner.
COD remittance tracking
Track COD collections from order-level expectation through remittance so delays are visible before they become stale follow-ups.
COD cycles are tracked across dispatch, delivery, remittance expectation, and bank credit so teams can see what is pending and what arrived.
AI review assistant
Ask why a record matched, which exceptions are blocking close, or what a fee variance actually means — answered against your live workspace, not a static export.
The assistant reads your current files, rule sets, and exception queue via tool-calling, so explanations cite real records and match decisions instead of generic advice.
AI-drafted match rules
Describe the reconciliation in plain English or let ReconPe infer rules from your data — no blank rule editor, no template hunting.
Every AI-generated rule is flagged for auditor traceability. Reviewers edit, accept, or override; rule sets stay versioned so teams know what a human signed off on.
Exception resolution suggestions
Each exception arrives with a proposed fix and the reasoning behind it, so analysts spend time deciding — not searching.
A second AI pass also surfaces probable matches that rigid rules missed — the kind finance teams usually catch only on a third read-through.
Exportable review reports
Generate review-ready reports with discrepancy context, status, and next-step information for finance and CA workflows.
Exports keep matched records, pending issues, and audit notes together so external reviewers do not need to rebuild the story manually.
Risk prioritization
Rank reconciliation output by operational risk so teams know what to review first.
ReconPe can weigh match quality, exception count, and financial impact to push the most important items to the top of the queue.
Review the highest-risk settlement items first.
ReconPe helps teams move beyond a flat matched or unmatched list by surfacing the discrepancies that need attention sooner.
- Prioritize records by impact and review urgency
- Add narrative context to exceptions and fee variance
- Separate clean matches from cases that need operator review
- Keep follow-up focused on business impact instead of file cleanup
Review note
"The current run shows commission variance across 12 Amazon India lines and delayed COD receipts in the most recent cycle. Review the fee breakup before closing the settlement."
Bring settlement control into one audit-ready workflow.
Start with 5 free reconciliations, then scale into commission audits, COD follow-up, and shared review workflows as volume grows.