24 prebuilt pipelines

Reconcile a bank feed against a ledger: normalize both, match with exact, amount-tolerance, date-proximity, and fuzzy-name strategies. Exceptions (≤5% of transactions) go to a human-review queue with per-item reasoning. Matched items write back to the ERP with read-back validation so the books stay the system of record.

Generate an accounts-receivable aging report from open invoices: bucket by 0-30, 31-60, 61-90, 91+ days, weighted-average days-sales-outstanding, and top-10 delinquent customers. The output is the artifact your collections team uses; the artifact's column-projection matches Big-4 workpaper templates.

Same shape as the AR aging but on the bills side. Buckets by due-date proximity, flags vendors above a configurable threshold (default $5K past 60 days), surfaces 1099-reportable vendors, and posts to the AP review queue.

Tie every balance-sheet account's GL balance to its sub-ledger or third-party confirmation. Output: one row per account with `gl_balance | subledger_balance | variance | variance_pct | due_to | next_step`. Used by audit teams as the workpaper tie-out.

Reconcile a credit-card statement: classify each charge as company/personal/uncategorized, split personal items to a reimbursement queue, post company items to the GL by category. Output: one row per charge with classification + GL account + receipt link.

Reconcile a Gusto 'Full Cash Requirements Report' to bank ledger: matches each pay-run and tax-deposit line, surfaces unmatched items as exceptions. Posts matched items to the payroll clearing account.

Identify and match intercompany transactions across related entities: same amount, same period, mirrored accounts. Surfaces unmatched pairs as exceptions for the consolidation team.

Compare financial metrics across periods: pivot amounts by account, surface variances above a configurable threshold (default 10% or $10K), classify variances as timing/permanent/one-time. Output: one row per account with current period, prior period, variance, variance %, classification.

Compare actual results to budget by account (and department if present). Output: standard budget-vs-actual table with variance and variance-pct, plus a footnote on classification.

Build a profit-and-loss statement segmented by department: revenue, COGS, opex, allocated shared costs, departmental contribution margin. Same shape as a standard departmental P&L; output is the artifact a multi-department CFO uses in board materials.

Consolidate multiple entities: combine balances by account, eliminate intercompany, apply currency translation if entities are in different functional currencies. Output: consolidated trial balance + elimination schedule + translation schedule.

Translate foreign-currency financials to the reporting currency using period-end or average rates per ASC 830 / IAS 21. Output: translated trial balance + CTA schedule.

Monthly inventory rollforward: opening stock → purchases → COGS → adjustments → closing stock. Output: one row per SKU with the four movements and the closing-units check (closing ≡ opening + net change).

Allocate a contract's transaction price across performance obligations (POs), recognize revenue as each PO is satisfied. Output: schedule of POs with `allocated_price | recognized_to_date | deferred | recognized_this_period`.

Compute tax depreciation for a fixed-asset register using the MACRS half-year or ADS straight-line convention. Output: schedule per asset with `class | method | life | basis | dep_this_year | accum_dep | nbv`.

Build the ASC 842 / IFRS 16 lease amortization schedule: PV of future lease payments, ROU asset, lease liability amortization, interest expense, operating expense. Output: one row per period with the four schedules.

Compute the current income-tax provision and effective tax rate: book income → permanent differences → temporary differences → taxable income → current tax. Output: provision waterfall with effective rate and reconciliation to the statutory rate.

Flag transaction-level anomalies via three independent statistical tests (Mahalanobis distance on amount+date, z-score on amount alone, isolation-forest on feature vector). Output: one row per flagged transaction with `anomaly_score | test_that_flagged | suggested_reason`.

Normalize a messy uploaded table at input: drop empty/noise rows, coerce types, normalize values, deduplicate. (Single-input pipelines only — per Aug 19 PR `23155cb6`, multi-input pipelines must label input count and refuse the cleansing chain.)

Apply consistent, professional Excel formatting to an existing deliverable: header-row bold, currency column as `#,##0`, frozen header, named ranges. Pure formatting — no data changes.

Find likely-duplicate vendor/customer master records (exact + fuzzy + phonetics). Output: cluster of records with `match_score | suggested_merge_target | merge_risk`.

Price metered usage against a versioned rate card (flat or volume-tiered). Output: one row per customer with `usage_quantity | tier_applied | unit_price | invoice_amount`.

Build fixed-rate, fixed-payment loan amortization schedules: payment allocation between principal and interest, ending balance per period. Output: one row per period.