Skip to content

Bridging (Direct & Indirect Forecasts)

Overview

Bridging refers to the process of reconciling and explaining the differences between direct cash flow forecasts (based on actual bank transactions and near-term receivables/payables) and indirect cash flow forecasts (derived from P&L budgets and balance sheet movements). This is a critical activity for treasury teams who need to explain variance drivers to management and ensure alignment between operational cash reality and financial planning.

For treasury teams, bridging enables confident communication with CFOs and boards about why actual cash differs from budgeted expectations. It reveals whether variances stem from timing differences, working capital movements, or fundamental forecast errors. Without effective bridging, teams spend excessive time manually investigating discrepancies and chasing local entities for explanations.

Palm can help by automating transaction classification, providing waterfall visualizations for variance analysis, and enabling drill-down into working capital components without manual Excel reconciliation.

For detailed ICP context and terminology, see fundamentals.md

Top Jobs & Desired Outcomes

Full history: jobs.md

1. Bridge direct and indirect cash flow forecasts to explain variance drivers ✓

Desired Outcomes: - Minimize the time required to identify deviations between direct cash flows and indirect budget - Reduce the frequency of having to ask local entities for variance explanations - Increase visibility into working capital movement drivers (AR vs AP breakdown) - Increase the ability to drill from consolidated variance down to entity and transaction level - Increase the ability to capture variance commentary directly against bridge line items - Persist confirmed variance explanations against past bridge periods (so the same investigation isn't repeated months later) - Reduce the number of hand-maintained assumption tables required to translate plan into cash - Decrease the structural complexity of the bridge by replacing rule chains with statistical inference

"For me, what I need to answer is always: what's the working capital deviation? What's driving that?" — Euroports (2026-04-15)

"There are two ways of telling if a forecast is good. First, the variance analysis... and the other one — we call it plausibility check — is, does my forecast more or less match what controlling is saying. This is what you're bringing here with data instead of a gut feeling." — ON (2026-04-29, Lucia)

Sources: Euroports (×2), Personio, Levi's, ON (×2) (Confirmed — 6 sources)

2. Provide timely feedback to FP&A and regional teams on actual cash performance ✓

Desired Outcomes: - Reduce the time to identify and communicate significant variances to FP&A and regional stakeholders - Increase the reliability of variance explanations with drill-down context that regional teams can act on - Increase treasury's ability to challenge regional teams when their forecasts are persistently off

"What I'm aiming to achieve is being able to use the system as like, a true reporting hub." — Personio (Tom)

"I think what's interesting here is for us to be able to go to the regional team and say, hey, your forecast is off — and we believe it's because your underlying P&L forecast." — ON (Lucia, 2026-04-29)

Sources: Personio, ON (Confirmed — 2 sources, promoted 2026-04-30)

⚡ Reduce misalignment between Treasury, FP&A, and Accounting categorizations

Desired Outcomes: - Minimize the effort to map Treasury cash flow categories to FP&A budget line items - Reduce the frequency of miscommunication caused by different teams using different category structures

"Our controlling team plans P&L and budget for the next five years. And then, as a consequence, they have some sort of cash planning. It's very difficult for me to tell if it's realistic." — ON (Lucia, 2026-03-04)

Source: ON (Emerging — single company, two conversations: 2026-03-04 + reinforced 2026-04-29)

Note: "Classify bank transactions into budget categories" is tracked in Categorization where it is a confirmed job.

Key Pain Points

Full history: pain-points.md

  • T+1 reporting takes hours of manual work - First day of month cash burn analysis (Source: Personio)
  • No central hub for variance analysis - Each analysis done differently (Source: Personio)
  • Manual plausibility check against the long-term plan is high-touch - Treasury sits in Excel/Anaplan checking how Palm's forecast compares to the long-term plan (Source: ON)
  • Anaplan bridge logic = high-maintenance assumption tables - AR aging splits, VAT timing, payment terms — re-justified each cycle (Source: ON)
  • Inventory→COGS→cash chain is structurally complex - Lucia's #1 candidate for replacing rule chain with statistical inference (Source: ON)
  • Black-box ML loses controlling/FP&A buy-in - Without explainability, treasury can't pitch Palm's logic as official (Source: ON)
  • Manual classification is time-consuming and error-prone (Source: Euroports)
  • Classification errors lead to "pollution of other categories" (Source: Euroports)
  • Current tools lack drill-down and variance analysis capabilities (Source: Euroports)
  • Manual splitting of FP&A monthly numbers into weekly forecasts (Source: Personio)
  • Difficulty separating payroll and supplier payments from statement data (Source: Personio)
  • Board deck manually rebuilt from Excel each quarter - copy tables + commentary into PowerPoint (Source: Euroports)
  • Free-text commentary only lives in Excel - no tool-native surface to explain variances against bridge rows (Source: Euroports)
  • Local entity forecast bias distorts numbers - AP overestimates payments, conservative entities match budget (Source: Euroports)
  • Factor payouts arrive as lump sums - no customer-level visibility into what drives factoring % changes (Source: Euroports)
  • Multiple ERPs (Oracle + InVision) complicate unified AR/AP ingestion (Source: Euroports)
  • Regional forecasts persistently off but treasury lacks a cash-bridge angle to challenge them - APEC region "always off because we grow faster than expected" (Source: ON)

Key Opportunities

  • Auto-classification with learning - Classify transactions automatically, allow corrections that retrain the model
  • Waterfall visualizations - Visual bridging from direct to indirect with drill-down
  • Working capital breakdown - AR vs AP movement visibility without manual analysis
  • In-product variance commentary with persistence - Capture commentary per row; surface prior comments when filtering past periods so the same investigation isn't repeated
  • Variance feedback loop into the model - User confirmations of variance drivers feed the model so future similar variances are pre-explained (parallel to categorization prompting)
  • Board-ready bridge reports - Generate the full quarterly bridge (Actual / Last Estimate / Previous Estimate / Prior Year / Budget) directly from Palm, no PowerPoint rebuild
  • Forecast accuracy scoreboard per entity - Drive accountability; reduce variance that has to be explained after the fact
  • Customer-level factoring drill-down - Explain working capital variance by identifying which customers drove factoring % changes
  • Replace rule-based bridge logic with model-driven inference - Long-term ambition: Palm ingests P&L + balance sheet, infers cash translation from historical patterns, and replaces hand-maintained Anaplan tables (ON's specific ask)
  • Explainability surface ("Gemini-style reasoning") - Show Palm's reasoning steps so treasury (and especially controlling/FP&A) can validate the logic — required to earn adoption as the official translation source
  • Boundary-testing modes - Accept inputs at three granularities (direct cash flow, indirect with same categories, raw P&L) so customers can find where the model's translation works and where it breaks (ON volunteered to pilot)

Open Questions

  • [ ] How do other companies handle factored receivables in their bridging process?
  • [ ] What's the typical frequency of bridging analysis? (Euroports does quarterly for board, entity-level more frequently)
  • [ ] Is the right v0 data path bank-statement categorization alone, or must AR/AP ingestion be Day 1?
  • [ ] Should commentary capture be per-row (Matthias' model) or per-variance-threshold (alert-driven)?
  • [ ] How do we reconcile multi-ERP realities (Oracle + InVision at Euroports) — file ingestion per source, or unified schema?
  • [ ] Can AI-generated variance commentary be trusted for board-level reporting, or only as a draft?
  • [ ] How does the bridge handle the time-horizon split (Palm = source ≤13wk, FP&A = source >13wk) in the UI? Single view with weighting, or two stacked views?
  • [ ] Where does the variance feedback loop live in the data model — per-period commentary tied to bridge rows, or feeding model retraining?
  • [ ] Can Palm replace rule-based assumption tables (AR aging, VAT timing) with statistical inference at v1, or is that a longer arc?
  • [ ] What's the minimum viable explainability surface to earn FP&A adoption — assumptions + calculation path, or full audit trail?
  • [ ] How do we represent boundary-testing modes in the product (direct / indirect / P&L inputs) without confusing the typical user?

Last updated: 2026-04-30 | Sources: 6 transcripts (view all)