ReconPe

0 reviews

ReconPe is an AI-native data reconciliation and financial-close platform for finance teams, e-commerce sellers, accountants, and CFOs. Upload source and target datasets in any common format (CSV, Excel, JSON, XML, TXT), let AI suggest field mappings and matching rules, and run audit-ready reconciliations — with every AI suggestion proposed for human review, never silently applied. At its core is an adaptive cascade engine that goes far beyond a simple VLOOKUP: exact-key, fuzzy text, token-based (free-text), keyless amount-and-date, and many-to-one aggregation matching (e.g. multiple invoices against one payment). Every match carries a confidence score — high-confidence matches auto-clear, and uncertain ones route into a typed exception queue with severity levels, role-based approval chains, and a full audit trail. One platform covers the full reconciliation and close lifecycle: • General reconciliation — reconcile any two datasets: bank vs books, AR vs GL, sub-ledger vs ledger, or any custom file-in → reconcile → done workflow. • Marketplace settlement reconciliation — for Indian sellers on Flipkart, Amazon, Meesho and 10+ marketplaces. Match payouts to bank credits, audit every order's commission against your rate card, detect fee leakage, validate GST on fees, and handle TCS (Sec 52) and TDS (Sec 194-O), plus real-world messiness like Indian comma formatting, truncated bank narrations, NEFT rounding, and multi-file aggregation. • GST reconciliation — match GSTR-2B against your purchase register, surface input tax credit (ITC) gaps, and tie out returns. • Tally Ledger Cockpit — import Tally Excel/CSV exports and reconcile customer, vendor, and bank ledgers with grouped matching and break classification. • FinanceOps close suite — budget-vs-actual variance with AI-narrated break drivers, prepaids and accruals, recurring journal entries, close automation and checklists, certification readiness, CFO close packs, and Tally XML export. Three ways to build and run the work: • Design Your Flow (Flow Studio) — a visual, no-code canvas to design reconciliation flows, describe matching logic in plain English, and run them as repeatable pipelines. • Marketplace Agent — reconciliation you program in plain English. Ask "did I get paid correctly on every order?" and it runs the engines, builds an action plan of what to recover, and remembers your business context across runs. • Fractional Bookkeeper Marketplace — connect with vetted bookkeepers and finance pros to run your books for you, scaling from self-serve to done-for-you. AI agents assist at every stage — mappings, exception investigation, counterparty patterns, narratives — under a strict AI-proposes, human-approves discipline. Multiple AI providers supported (Anthropic, OpenAI, DeepSeek, Gemini), configurable per organization. Outputs are audit-ready: reconciliation certificates, exception reports, lineage, and close documentation — built for teams that need to defend their numbers.