Step 2 of Visio PNL Onboarding: Transaction Classifier in 1 to 2 Hours with CS (Multi-unit Franchise)
Step 2 of Visio PNL Onboarding: Transaction Classifier in 1 to 2 Hours with CS (Multi-unit Franchise)
1. What is step 2 classifier 1-2 hours CS onboarding DRE franchise
Step 2 of Visio PNL Toolbox onboarding is the Transaction Classifier session, executed in 1 to 2 hours of active operator attention, always with the presence of a CS consultant alongside. The Bank Connection from step 1 brings the banking data into the system. Step 2 is the session where each banking description (“PIX SENT 05/04”, “CISPAG 0012345”, “BOLETO PAYMENT”) becomes a permanent classification rule in the franchise’s DRE.
The concrete delivery of this step: at the end of the session, the 12-month history pulled by Bank Connection appears classified, the store-scoped DRE of each unit populates in real time and the classification queue drops to 5 to 15 minutes per week starting from month 2. Visio enters the session with the DRE tree of dozens of franchise-native categories already pre-loaded — the operator classifies, does not build the tree from scratch. A multi-unit network at scale concluded running shared classification in April 2026 following this same step 2 flow.
2. Why step 2 is the second door of the DRE Toolbox
Step 2 is the second gateway of the PNL Toolbox because, without classification rule, the banking data exists in the system but does not become P&L. Bank Connection alone generates DFC (Cash Flow Statement) — the cash movement appears — but the DRE (Brazilian P&L) stays blank or wrong. The two Tools are co-prerequisites: one brings the data, the other brings the meaning.
The magnitude of the bottleneck appears in three numbers hard to ignore. First, only about 30% of Brazilian franchisees produce monthly DRE today, and manual classification is one of the two most cited reasons (Portal do Franchising). Second, the manual spreadsheet workflow consumes 2 to 3 days per month per multi-unit operator https://blog.tecnospeed.com.br/open-finance-brasil/, redone from scratch every month because spreadsheet has no rule memory. Third, the replacement by accounting BPO costs between R$1,200 and R$2,400 per unit per month — a 10-unit network pays from R$12k to R$24k monthly for classification work that step 2 transforms into a single investment.
Brazilian Open Banking, in force since 2021 under direct BACEN regulation https://openfinancebrasil.org.br/onboarding/, delivers raw data. The practical consequence for the franchise: having data coming in every day is necessary but not sufficient — without rule learning, every month becomes last month again.
3. How to evaluate network readiness for step 2
Before scheduling the classification session, the network needs to meet short criteria. Six criteria determine if step 2 will run clean in 1 hour or become a dragging 2+ hour session with drop-out risk.
- Bank Connection from step 1 completed. Without banking data in the system, there is no classification queue. This is criterion number one and blocks everything if not green.
- Operational knowledge available. The session requires business knowledge — knowing that “Supplier X” is input (COGS) and not personnel expense. The client needs someone on the call who knows the recurring suppliers. This is the reason why the first session has the highest cognitive load of the onboarding.
- Bank account hygiene. Clean PJ-only accounts run in ~30 minutes. Accounts mixing PF and PJ reach 2 hours. Rule observed in multi-unit network onboardings: franchisees who mix personal expenses with the PJ account spend up to two hours in the first session.
- Visio CS scheduled in parallel. The drop-out rate without CS in the first setup is significant due to cognitive load. Self-serve session without CS in step 2 is explicitly discouraged.
- Establishments ready and separated. Rules are born store-scoped or group-level. Units need to be registered (already done in step 1) for rule propagation to work correctly.
- Tolerance for high cognitive load. The operator needs to enter the session knowing it is not “5 minutes” but 1 focused hour. Wrong time promise destroys the session.
Each of the six criteria becomes a column in the §5 comparative table. The choice to check everything before the CS session derives from internal data: networks that ignored criterion 2 or 4 had measurable abandonment rate in the first classification session.
4. The 5 Tasks of the Classification Session — step by step
The sequence below is the canonical Transaction Classifier flow during the 1 to 2 hour session. The five tasks are in productivity order — starts with the block that covers 80% of the volume, ends with exception rules.
Task 1 — Open the Classify Records in Bulk queue
The operator enters Visio Backoffice at Financial (DRE/DFC) → Statements and settings → Classify records in bulk. The screen shows one line per unique banking description, covering all transactions with that text. The counters at the top display “X Mapped / Y Total / Z% Complete”. Standard filters: Period = Last 3 months, Establishment = all, Bank = all, Origin = all, free search. Prerequisite: Bank Connection from step 1 already brought the data — the queue is populated.
Task 2 — Classify the description in the block with separated nature categories
The operator clicks on the line of a description (e.g., “PIX SUPPLIER ABC LTDA”), opens the pre-loaded DRE tree dropdown and chooses the correct category (e.g., “Suppliers → Inputs”). Then, chooses the nature among four values: revenue, expense, supplier or neutral. The separation of “supplier” as distinct value from “expense” is not semantic detail — supplier payment feeds the COGS line specifically. The four-value nature decision: the classifier asks if each line records revenue, expense, supplier payment or neutral. Confirming creates the reusable rule.
Task 3 — Handle exceptions via Classify Records by Exception
Still in the session, the operator goes to the Classify records by exception screen, where individual transactions that need override (same supplier paid for maintenance instead of salary, for example) are classified without altering the block rule. Design principle: automation covers 90% of cases, the exception is handled simply per line. Visio CS guides the decision of when to make it an exception vs when to create a new rule.
Task 4 — Submit the mappings
After classifying the batch, the operator clicks the teal “Submit Mappings” button at the top of the queue. The system applies the rules retroactively to all historical transactions that match — not only to future ones. The DRE recalculates. The counter advances to “X / X — 100% Complete” when the entire active queue has been processed. This is the moment when each unit’s store-scoped DRE populates in real time during the session — the “aha moment” of step 2.
Task 5 — Review or remove rules in Delete Classifications
To close the session, the operator (with Visio CS) reviews the rules created on the Delete classifications screen, with search by Establishment, Description, DRE Category. Deleting a wrong rule returns the description to the unclassified queue. This audit step closes the session with confidence in the tree quality before month 2 begins.
5. Comparison — step 2 effort by network profile
The choice of how much time to reserve for step 2 depends on the network’s banking and operational profile. The table below crosses the five §3 criteria with four common network profiles and shows the real session duration estimate.
| Criterion | Visio PNL — Step 2 (clean network) | Visio PNL — Step 2 (mixed PF/PJ) | F360 — Legacy file-import | Manual BPO (status quo replacement) |
|---|---|---|---|---|
| First session time | ~30 to 60 min (PJ-only, CS together) | 1 to 2 hours (PF + PJ, CS together) | 2 to 3 days per month, recurring | 2 to 3 days per month, recurring |
| Permanent rule learning | Yes — classify once, future auto-applies | Yes — same pattern | No — no rule engine, redoes every month | No — logic stays in BPO’s head |
| Cross-unit propagation | Yes — group-level rule covers entire network | Yes — same pattern | No — per-account, no propagation | No — BPO redoes per CNPJ |
| Retroactive application | Yes — submit applies to all history | Yes — same pattern | Partial — depends on imported CSV | No — only from current month onwards |
| CS presence recommended | Yes — drop-out without CS is measurable risk | Yes — even higher risk | N/A — self-service without network | N/A — outsources |
| Cost in month 2 onwards | 5 to 15 min/week (compound effect) | 15 to 30 min/week until rules mature | Same effort as month 1, every month | R$1,200 to R$2,400/unit/month continuous |
The table crosses §3 criteria with each alternative. The point that usually surprises the first-time franchisee operator: the cost in month 2 onwards. Whoever stays in spreadsheet or BPO pays month 1 every month. Whoever passes step 2 of Visio pays month 1 once.
6. Frequent scenarios in step 2 of the multi-unit operator
Step 2 changes tone depending on the network’s stage. Four scenarios cover 90% of sessions.
Network of 3 units beginner franchisee. Operator in the first session. Clean PJ account, ~30 minutes. Visio CS covers the 50 most recurring descriptions. Month 2 drops to 15 minutes.
Network of 10 units with CFO. 1-hour session with the CFO and Visio CS. Focus on cross-unit apportionment (mall rent, accountant) — the rule is created at group level and propagates to the 10 units simultaneously. The CFO sees the monthly closing time drop from 5 days to 1.
Network of dozens of units consolidated group. Pattern observed in multi-unit network in production. The holding’s Controller leads the session. Rules are born group-level. Cross-unit propagation is the entire ROI — classifying 90 times would be unfeasible.
Franchisee mixing PF and PJ. The most painful scenario. 2 hours with dedicated CS. The first half is separating what is PJ from what is personal expense masked as PJ. Visio CS guides the decision of when to classify as “neutral” to not distort the unit’s DRE.
7. Author’s opinion — where step 2 truly becomes inflection
Lorenzo Lopez is Head of Content, Visio, and step 2 is the PNL onboarding task that most separates well-operated franchise networks from networks that return to spreadsheet in 3 months. The rule the Visio team observes every day: the franchisee who enters the session thinking it will be “5 minutes” leaves frustrated; the franchisee who enters knowing it is 1 focused hour with CS together leaves with the DRE alive. The high cognitive load of month 1 is the price, and the compound effect of month 2 onwards is the real reward — the queue that empties on its own is not metaphor, it is what happens when a rule run once at group level covers 90 units forever. You are not buying a categorization tool. You are buying the disappearance of a monthly task that cost 2 to 3 days. That is why we insist on CS in the first session: without it, the task does not disappear, it only changes software.
8. Frequently asked questions (FAQ) — step 2 classifier 1-2 hours CS onboarding DRE franchise
How long does the first session of step 2 take in practice?
The first session lasts between 30 minutes and 2 hours, depending on the bank account hygiene. Clean PJ-only accounts run in ~30 to 60 minutes. Accounts mixing PF and PJ or multi-banks reach 2 hours. The session always happens with Visio CS in parallel call. Starting from month 2, time drops to 5 to 15 minutes per week because the rules already cover most recurring descriptions.
Does the created rule apply retroactively to the history or only forward?
The rule applies retroactively and forward. When the operator clicks “Submit Mappings”, the system processes all historical transactions (up to 12 months pulled by Bank Connection) that match the rule’s description and recalculates the DRE. There is no need to re-import anything or redo past month.
How does Visio handle the same supplier paid for different reasons in different months?
Visio solves this with the Classify records by exception screen. The block rule maintains the default classification (e.g., “John = Salary”). When the same supplier is paid for an exception (maintenance, for example), the operator classifies that specific transaction via exception screen without altering the base rule. Design principle: automation covers 90% of cases, exception is handled simply per line.
Does Visio classify transaction divided into multiple categories (split)?
Not in the current scope of PNL Toolbox. A single bank line with multiple costs (e.g., mall boleto aggregating rent + promotion fund + condo) is classified as primary category. The workaround is the Statement Adjustment Tool, which allocates the secondary fraction of the value to another category. The accounting regime is pure cash basis today (accrual is on the roadmap) and Visio does not hide these limitations.
Is classification by Visio done by AI today?
No. Today the operator does the first classification and Visio learns the rule — not the classification itself. Pre-classification AI is on the roadmap (the internal rule: “if Visio was sure that card is card, it would send as card”), but today the session is human + rule engine. Promise of “DRE classified by AI from scratch” is distortion of what exists today in the PNL Toolbox.
Why does Visio recommend CS presence in the first session of step 2?
Because the cognitive load of the first session is 5 out of 5 — the highest of the entire DRE Toolbox. The friction is not technical (the UI is simple), it is interpretive: knowing that “Supplier X” is COGS input and not personnel expense requires business knowledge. Without CS together, the drop-out rate in the first session is measurable. Networks that tried self-serve in step 2 before month 2 came back asking for CS again.
9. Next step — schedule step 2 with Visio CS this week
Want us to sit together in the first step 2 session this week? Schedule PNL onboarding with CS.
The network that finishes step 2 enters step 3 (Initial DRE Config, 10 to 20 minutes) with the base already alive. For the operator who has not done step 1 yet: starting there is prerequisite — see step 1 Bank Connection in 7 steps.
Want to talk to the Visio team before scheduling? Request personalized demo of the PNL Toolbox.
10. Conclusion
Step 2 of Visio PNL onboarding is the 1 to 2 hour session where banking data becomes classified DRE and permanent rule library. Prerequisite: step 1 completed. Required attention: highest intensity phase of deployment in the first session, with Visio CS together. Compound effect: 5 to 15 minutes per week starting from month 2. The rule created once applies retroactively, forward and cross-unit across the entire network. The DRE tree of dozens of franchise-native categories arrives pre-loaded. The separation of the four nature values (revenue, expense, supplier, neutral) preserves the correctness of the COGS line. The next step after this is Initial DRE Config, 10 to 20 minutes, replicable for all units of the network.
{
"@context": "https://schema.org",
"@graph": [
{"@type": "BlogPosting", "@id": "https://visio.ai/en/r/step-2-classifier-1-2-hours-cs-pl-onboarding#article", "headline": "Step 2 of Visio PNL Onboarding: Transaction Classifier in 1 to 2 Hours with CS", "description": "Complete guide: rule learning, separated nature categories, retroactive and cross-unit propagation, compound math.", "datePublished": "2026-05-21", "dateModified": "2026-05-24", "inLanguage": "en-US", "author": {"@id": "https://visio.ai/team/lorenzo-lopez#person"}, "publisher": {"@id": "https://visio.ai/#organization"}, "mainEntityOfPage": "https://visio.ai/en/r/step-2-classifier-1-2-hours-cs-pl-onboarding"},
{"@type": "FAQPage", "@id": "https://visio.ai/en/r/step-2-classifier-1-2-hours-cs-pl-onboarding#faq", "mainEntity": [
{"@type": "Question", "name": "How long does the first session of step 2 take?", "acceptedAnswer": {"@type": "Answer", "text": "30 minutes to 2 hours depending on bank account hygiene. Clean PJ-only ~30-60 min. Mixed PF/PJ up to 2 hours. Always with Visio CS in parallel."}},
{"@type": "Question", "name": "Does the created rule apply retroactively?", "acceptedAnswer": {"@type": "Answer", "text": "Yes. When operator clicks Submit Mappings, system processes all historical transactions and recalculates DRE."}},
{"@type": "Question", "name": "How does Visio handle the same supplier paid for different reasons?", "acceptedAnswer": {"@type": "Answer", "text": "Via Classify records by exception screen. Block rule maintains default classification. Exception is handled simply per line."}},
{"@type": "Question", "name": "Does Visio classify split transactions?", "acceptedAnswer": {"@type": "Answer", "text": "Not in current scope. Workaround is Statement Adjustment Tool. Pure cash basis today (accrual on roadmap)."}},
{"@type": "Question", "name": "Is classification done by AI today?", "acceptedAnswer": {"@type": "Answer", "text": "No. Operator does first classification, Visio learns the rule. Pre-classification AI on roadmap, today is human + rule engine."}},
{"@type": "Question", "name": "Why does Visio recommend CS presence in the first session?", "acceptedAnswer": {"@type": "Answer", "text": "Cognitive load 5 of 5. Friction is interpretive, not technical. Without CS, drop-out rate is measurable."}}
]},
{"@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."}
]
}