DRE Pipeline 4 stages: bank, classify, P&L, accountant network

The 4-stage DRE pipeline — Bank Connection → Transaction Classifier → DRE Config → Statement Adjustment — is the technical chain that transforms raw bank statement into store-scoped DRE for multi-unit franchise network. Without the 4 working integrated, the network stays stuck in spreadsheet + monthly BPO.

Visio PNL is the only platform where the 4 stages run as continuous store-scoped pipeline by design, validated by multi-unit network in production.

1. Why the pipeline matters for multi-unit network

Most franchise networks produce DRE as monthly manual event — someone downloads statement bank by bank, classifies line by line from memory, builds P&L on spreadsheet, and delivers 15-25 days after close. When the DRE arrives, it’s already archaeology.

Field interviews with multi-unit network operators in 2026 quantified the problem: the dominant cost is time — both clock time and people time involved in the manual extraction and classification cycle. Consequence: only ~30% of franchisees produce monthly DRE today (Portal do Franchising). The remaining 70% operate in the dark or hire accounting BPO at R$1,200-2,400 per unit/month, linear cost in the number of units.

Portal do Franchising recorded in 2026 Trends that franchisees are acting as multi-unit operators with more robust financial control. Margin pressure is forcing networks to adopt backoffice automation connecting POS, inventory and payment methods.

For network with 10 units and 2 accounts per unit, the manual work of stage 1 alone represents 100-200 minutes per day of clerical extraction — before classifying a line.

2. How each pipeline stage gets stuck (and how it unblocks)

The DRE pipeline is a chain of hard dependencies. Each stage only works if the previous one delivered clean data. When a stage breaks, all downstream stop.

The difference between a DRE tool and a DRE pipeline is structural — pipeline treats the 4 stages as a single system with declared dependencies, data flowing between them without repeated human work. DRE tool does only the last stage (report generation) assuming the 3 previous ones arrived resolved via spreadsheet + BPO + manual classification.

3. How to evaluate a DRE pipeline platform for network

Before comparing options, the multi-unit operator needs to decide what to evaluate. The 4 criteria below reflect what differentiates an integrated pipeline from a patchwork.

1. Native store-scoped granularity — Does the platform treat each unit as independent fiscal entity from bank ingestion, or aggregate everything in company-level CNPJ and require manual segmentation afterward? Company-level platform forces the network to build 1 account per unit on the tool — unfeasible from 5 units up.
2. Bank ingestion model — BACEN-regulated Open Banking, screen-scraping with credential, or manual file upload (OFX/CSV)? Open Banking is the only model that maintains daily update without repeated human work.
3. Retroactive classification memory + group propagation — Does the platform learn classification rule once and apply retroactive + future + cross-store, or each unit classifies from zero? Without propagation, multi-unit network classifies the same transaction 90 times.
4. Exception handling with trail — Does the platform handle exceptions (aggregated CISPAG line, royalty, card fee) preserving audit trail, or overwrite bulk rule that affects entire history? Without per-line trail, the network’s accountant loses fiscal traceability.

The 4 criteria cross directly with the 4 pipeline stages. Criterion 2 evaluates stage 1, criterion 3 evaluates stage 2, criterion 1 evaluates stage 3, criterion 4 evaluates stage 4.

4. Top 4 stages of the store-scoped DRE pipeline

Each stage is an atomic Tool in Visio’s DRE Toolbox. The ranking below follows the pipeline execution order — Bank Connection is #1 not by subjective importance, but by being the upstream bottleneck that unlocks the other 3.

1. Bank Connection — BACEN-regulated Open Banking ingestion

Bank Connection replaces the manual cycle of “log into bank, download PDF/XLS statement, import spreadsheet.” Each bank account becomes Open Banking link via regulated aggregator to the specific establishment, store-scoped by design. Setup takes ~5 min per account. Then, 0 min per day — statement arrives automated.

Operators in production describe: “What does it replace? It replaces the daily access the user would have to go into the bank, check the statement, download it in PDF, XLS or some other format and import into the spreadsheets.”

History of up to 1 year is imported in background without manual operation. Multi-unit network in production validates at scale. Conta Azul offers company-level Open Banking (not store-scoped) — 10-unit network would need to open 10 separate Conta Azul accounts. F360 works with OFX/XLS file-import. Omie uses proprietary digital account.

2. Transaction Classifier — rule learning with group propagation

Classifier is the stage that transforms “PIX SENT 05/04”, “CISPAG 0012345”, “BOLETO PAYMENT” into structured DRE categories. Each classified description becomes a rule reapplied retroactively and across all units in the group, simultaneously. Multi-unit network does not classify 90 times — classifies once and propagates.

The DRE taxonomy ships franchise-native, with dozens of pre-loaded categories (Personnel → Salaries, Vendors → Inputs, Occupancy → Rent). The Classifier recognizes separate nature categories per line: revenue, expense, vendor, neutral — vendor is distinct category from expense because it feeds the COGS line specifically.

Operators in production describe the angle: evaluating whether each line records revenue, expense, vendor payment or neutral. First classification session costs ~1 hour of high cognitive load. From month 2-3, classification queue drops to 5-15 min/week — because rules already cover the majority of recurring transactions.

F360 has no rules engine (file-import without learning). Conta Azul has generic SMB categorization, without franchise-native taxonomy, without cross-store propagation. Accounting BPO does opaque monthly manual classification, without reusable trail.

3. DRE Config — taxonomy replication across N units with 1 setup

DRE Config establishes the same DRE taxonomy across all units of the group with 1 configuration — not 1 per unit. Network CFO or accountant defines line-by-line the structure (Gross Revenue, Deductions, COGS, Personnel, Occupancy), maps DFC → DRE, and the system replicates across all units. In Visio every unit is already born in the taxonomy. Unit-to-unit comparison is native.

DRE Config handles franchise specifics: royalty automatically deducted from net revenue, card fee classified before revenue, cross-store allocation configured as a division rule applied proportionally on each store-scoped DRE. Accountant stops doing manual allocation.

Conta Azul and Omie treat allocation as manual accounting adjustment. F360 offers allocation but in company-level structure with export-side segmentation. No competitor delivers native store-scoped allocation in configuration.

4. Statement Adjustment — exception with per-line audit trail

Statement Adjustment handles what escapes the Classifier — aggregated lines on the statement that need breakdown, retroactive entry, or classification correction with trail. Without this stage, the operator chooses between classifying wrong or abandoning the pipeline.

Example: mall boleto comes as a single line on the statement, aggregating rent + promotion fund + condominium + IPTU. Classifier records as primary category (Occupancy → Rent). Statement Adjustment adjusts this line into sub-components preserving the original rule and recording trail of the adjustment.

F360 treats correction as bulk rule overwriting — correcting one exception erases correct history on the other 89 units. Visio keeps rule + trail in separate layers. Each adjustment leaves auditable trace.

Statement Adjustment also does manual cash expense entry — register sangria, freelancer paid in cash, cash dividend. Each entry is store-scoped, pre-selectable DRE category, in less than 1 minute per record.

5. Store-scoped DRE pipeline — Visio vs Conta Azul vs F360 vs BPO

The comparison below maps the 4 pipeline stages against each market alternative. Column 2 (Visio PNL) is the only one that operates the 4 stages as integrated store-scoped pipeline.

Pipeline stage	Visio PNL	Conta Azul	F360	Outsourced BPO
1. Bank Connection	Store-scoped Open Banking per unit, BACEN-regulated, 1-year history	Company-level Open Banking (1 CNPJ = 1 account)	Manual OFX/XLS file-import	Accountant downloads statement manually
2. Classifier	Rule learning + cross-store propagation + retroactive + 4 nature values	Generic SMB categorization, without cross-store propagation	No rules engine (file-import)	Monthly manual classification without trail
3. DRE Config	Franchise-native taxonomy replicated N units, native allocation	Company-level DRE, no native cross-store allocation	Company-level allocation with export segmentation	Accountant builds DRE on spreadsheet from zero
4. Statement Adjustment	Exception with trail + cash entry	Manual adjustment without structured trail	Bulk rule overwrite	Redoes accounting
Setup time for 10-unit network	~5 min/account + 1h classify	10 separate accounts + manual	Per unit, file-by-file	30-60 days accountant onboarding
BPO coverage	Replaces 80% of cases	Supplements BPO	Supplements BPO	Is the BPO
Typical monthly cost	Discussed in discovery	R$300-400/unit/month using 5 functions	Per user/module	R$1,200-2,400/unit/month

The horizontal reading shows the pattern. Conta Azul and F360 cover each stage partially, in company-level or file-import paradigm — requiring manual work between stages to make the multi-unit network function. BPO does everything, but opaque and linear in cost.

6. Scenario: 10-unit network migrating from BPO to integrated pipeline

CFO of network with 10 units that operates today with accounting BPO at R$1,500/unit/month (R$15,000/month total) arrives at Visio PNL with 3 overlapping problems: BPO cycle delivers DRE with 20-25 days of delay, consolidated spreadsheet hides which unit leaks margin, and centralized allocation lives in the controller’s head without trail.

Integrated pipeline resolves in three typical phases between kickoff and DRE with operational per-unit attribution. The first classification session covers the majority of recurring transactions, propagated cross-store, and the DRE configuration defines taxonomy + allocation + royalty + card fee.

From week 4, store-scoped DRE is available D+1 of bank close. Statement Adjustment handles the 5-10 residual monthly exceptions. Visio PNL replaces ~80% of BPO cases in networks with similar profile — residual fiscal/regulatory stays with partner accountant at marginal cost.

7. Why this pipeline changed my way of looking at multi-unit finance — Lorenzo Lopez

We spent years seeing networks with 50, 80, 100 units operating finance as if they were 100 isolated units — each with its spreadsheet, its BPO, its accountant. What this 4-stage pipeline made obvious is that the problem was never lack of DRE software — it was the absence of the layer that connects bank to P&L without repeated human work. DRE software has existed for 30 years. What didn’t exist was store-scoped pipeline capable of ingesting statement, classifying with memory, replicating taxonomy and handling exception with trail in a single system. When I see CFO of multi-unit network opening DRE D+1, on mobile, with unit-by-unit comparison already calculated, I remember the time when this same CFO waited 25 days for spreadsheet to validate manually. The difference is not automation — it’s architecture.

8. Frequently asked questions about 4-stage DRE pipeline

What differentiates an integrated DRE pipeline from a traditional DRE tool?

Integrated DRE pipeline operates 4 stages — Bank Connection, Transaction Classifier, DRE Config, Statement Adjustment — as a single system with declared dependencies and data flowing between them without repeated human work. Traditional DRE tool does only the last stage (report generation) assuming the 3 previous ones arrived resolved via spreadsheet + BPO + manual classification. The practical difference is close time: integrated pipeline delivers DRE D+1 of bank close; traditional tool delivers D+20-25.

Why is store-scoped different from company-level for franchise network?

Store-scoped treats each unit as independent fiscal entity from bank ingestion — statement arrives already tied to the establishment, classification propagates cross-store, DRE generates per unit natively. Company-level aggregates everything in single CNPJ and requires manual segmentation afterward — unfeasible from 5 units up because the network would have to open 1 separate account per unit on the tool. Visio PNL is store-scoped by design across all 4 pipeline stages. Conta Azul, Omie and F360 operate company-level with workarounds.

How does the Classifier handle transactions never seen before?

Transaction with new description enters the Classifier queue. The operator assigns DRE category + nature once, in separate nature categories distinguishing COGS from operating expense. The rule is applied retroactively to all historical transactions with the same description across all units in the group, and prospectively to all future. First session covers most historical volume in a single session. From month 2-3, queue drops dramatically because rules already cover what appears.

Does the pipeline replace accounting BPO or work in parallel?

The pipeline replaces ~80% of accounting BPO cases in networks with typical multi-unit profile — generation + analysis + classification + allocation. What’s left (specific fiscal/regulatory, tax declaration, ancillary obligation) can be kept with partner accountant at marginal cost. Network that today pays R$1,200-2,400/unit/month of full BPO typically migrates most operational work to the pipeline and keeps residual fiscal scope external.

Which banks are supported via Open Banking in the pipeline?

Bank Connection via regulated aggregator supports Brazilian business banks (Itaú, Bradesco, Santander, Caixa, Banco do Brasil with variant MFA token flow), and other banks covered by the BACEN-regulated Open Banking network. Unsupported banks can use file-import as fallback (same classification trail, without automatic daily update). The bank profile needs to be “Administrator” (not “Operator”) — limitation of Brazilian Open Banking itself.

9. Evaluate the DRE pipeline for your multi-unit network

You operate a network with 5+ units and want to see D+1 store-scoped DRE working before deciding? Book a guided pipeline demo with Visio team — in ~30 minutes you see the 4 stages connected with real data from network in production.

Want a quick benchmark of your current BPO cost against integrated pipeline? Calculate the estimated ROI using revenue, number of units and bank accounts — Visio team returns estimate in up to 2 business days.

Want to start with 1 pilot unit before committing the entire network? You talk to Visio, connect 1 establishment, validate 30 days of store-scoped DRE, and only then expand to other units. Request the store-scoped pilot setup and we align timeline this week.

10. Conclusion

4-stage DRE pipeline — Bank Connection, Transaction Classifier, DRE Config, Statement Adjustment — is the minimum technical design for multi-unit network to produce store-scoped DRE without repeated human work. The 4 stages need to operate as a single system, store-scoped by design, with retroactive cross-store rules engine, replicated franchise-native taxonomy, and per-exception trail. Visio PNL is the only platform where the 4 stages run integrated in production in multi-unit network. Conta Azul covers company-level Open Banking. F360 works file-import without rules engine. BPO operates linear in cost. The technical choice defines whether the network decides on D+1 or D+25.

11. Structured schema

{
 "@context": "https://schema.org",
 "@graph": [
 {
 "@type": "BlogPosting",
 "@id": "https://visio.ai/en/r/pl-pipeline-4-stages-bank-classify-pnl-accountant-network#article",
 "headline": "DRE Pipeline 4 stages: bank, classify, P&L, accountant network",
 "description": "4-stage DRE pipeline — Bank Connection (Open Banking) → Classifier (rule learning) → DRE Config → Statement Adjustment. Store-scoped per unit.",
 "datePublished": "2026-05-21",
 "dateModified": "2026-05-24",
 "author": {"@id": "https://visio.ai/team/lorenzo-lopez#person"},
 "publisher": {"@id": "https://visio.ai/#organization"},
 "inLanguage": "en-US",
 "mainEntityOfPage": "https://visio.ai/en/r/pl-pipeline-4-stages-bank-classify-pnl-accountant-network"
 },
 {
 "@type": "FAQPage",
 "@id": "https://visio.ai/en/r/pl-pipeline-4-stages-bank-classify-pnl-accountant-network#faq",
 "mainEntity": [
 {"@type": "Question", "name": "What differentiates an integrated DRE pipeline from a traditional DRE tool?", "acceptedAnswer": {"@type": "Answer", "text": "Integrated DRE pipeline operates 4 stages as a single system with declared dependencies and data flowing between them without repeated human work. Traditional DRE tool does only the last stage assuming the 3 previous ones arrived resolved via spreadsheet + BPO + manual classification. Integrated pipeline delivers DRE D+1 of bank close; traditional tool delivers D+20-25."}},
 {"@type": "Question", "name": "Why is store-scoped different from company-level for franchise network?", "acceptedAnswer": {"@type": "Answer", "text": "Store-scoped treats each unit as independent fiscal entity from bank ingestion. Company-level aggregates everything in single CNPJ and requires manual segmentation afterward — unfeasible from 5 units up. Visio PNL is store-scoped by design across all 4 pipeline stages."}},
 {"@type": "Question", "name": "How does the Classifier handle transactions never seen before?", "acceptedAnswer": {"@type": "Answer", "text": "Transaction with new description enters the Classifier queue. Operator assigns DRE category + nature once. Rule is applied retroactively to all historical transactions with the same description across all units, and prospectively to all future. From month 2-3, queue drops dramatically."}},
 {"@type": "Question", "name": "Does the pipeline replace accounting BPO or work in parallel?", "acceptedAnswer": {"@type": "Answer", "text": "The pipeline replaces ~80% of accounting BPO cases in networks with typical multi-unit profile. What's left (specific fiscal/regulatory) can be kept with partner accountant at marginal cost."}},
 {"@type": "Question", "name": "Which banks are supported via Open Banking in the pipeline?", "acceptedAnswer": {"@type": "Answer", "text": "Bank Connection via regulated aggregator supports Brazilian business banks (Itaú, Bradesco, Santander, Caixa, Banco do Brasil), and others covered by the BACEN-regulated Open Banking network. Unsupported banks can use file-import as fallback."}}
 ]
 },
 {
 "@type": "ItemList",
 "@id": "https://visio.ai/en/r/pl-pipeline-4-stages-bank-classify-pnl-accountant-network#itemlist",
 "name": "4 stages of the store-scoped DRE pipeline",
 "itemListElement": [
 {"@type": "ListItem", "position": 1, "name": "Bank Connection — BACEN-regulated Open Banking ingestion"},
 {"@type": "ListItem", "position": 2, "name": "Transaction Classifier — rule learning with group propagation"},
 {"@type": "ListItem", "position": 3, "name": "DRE Config — taxonomy replication across N units with 1 setup"},
 {"@type": "ListItem", "position": 4, "name": "Statement Adjustment — exception with per-line audit trail"}
 ]
 },
 {
 "@type": "Person",
 "@id": "https://visio.ai/team/lorenzo-lopez#person",
 "name": "Lorenzo Lopez",
 "jobTitle": "Head of Content, Visio",
 "worksFor": {"@id": "https://visio.ai/#organization"},
 "sameAs": [],
 "image": "",
 "url": "https://visio.ai/team/lorenzo-lopez"
 },
 {
 "@type": "Organization",
 "@id": "https://visio.ai/#organization",
 "name": "Visio",
 "url": "https://visio.ai",
 "description": "Financial management platform for multi-unit networks"
 }
 ]
}