Cash Flow Copilot
An ongoing financial assistant on WhatsApp that forecasts cash inflows, flags shortfalls and helps managers decide what to do next.
Cash Flow Copilot is the core product. Partner Workspace brings it to companies through their advisors. Cash Flow Diagnostic gives companies a free starting point.
An ongoing financial assistant on WhatsApp that forecasts cash inflows, flags shortfalls and helps managers decide what to do next.
A workspace for accountants and advisory firms to manage their client portfolio and work within each company's financial context.
An automated diagnostic using ERP data to compare scheduled receivables with likely collections and quantify the gap.
Product definitions and commercial terms from Designing the product for GTM, pp. 1–5 and WOW Pitch 2026.2 (EN), p. 8. These descriptions do not establish current product availability.
The GTM plan defines three distinct roles: deliver recurring value, distribute the solution through partners and convert the initial experience into a subscription.
Anticipate cash shortfalls and support day-to-day financial decisions.
A subscription to an ongoing financial assistant combining document capture, reconciliation and predictive intelligence. Liquidity monitoring and decision support are part of the same offering.
Kaigoo does not transfer funds, provide credit or act as a credit intermediary. Rates are indicative; companies negotiate with their chosen financing provider.
C01–C03 · monitoring, alerts and conversation · C04–C06 · capture, approval and outcome
J3 · understand cash flow and act · J4 · submit, review and post
GTM · pp. 3–4, 6–7 · Pitch EN · pp. 4–6, 8 · Forecasting Engine · p. 1
Help partners serve more companies with the same team.
A workspace for managing a partner's client portfolio. It brings client companies together and provides access to each company's diagnostic and financial workflows, distributing the Copilot through accounting and advisory firms.
The materials target up to 40 companies per analyst. This is an ambition, rather than measured productivity or a contractual limit. Prioritization, filters and permissions require specifications to deliver the planned operation.
Expose the gap between cash scheduled in the ERP and cash likely to be collected.
An automated report using ERP data to compare receivables recorded at face value with forecast collections based on payment behavior. It exposes the cash that appears on the receivables schedule but may not arrive as expected, before the company subscribes to ongoing monitoring.
The local implementation compares receivables per ERP app connection. Without bank balances and accounts payable in the calculation, its result does not represent the company's full net cash position. D01 and D03 cover fields, data coverage and presentation.
R01–R03 · entry, diagnostic and subscription path
J1 · review the diagnostic · D01 · report content · D03 · data coverage · D02 · conversion
Free Cash Flow Diagnostic to uncover the issue; Cash Flow Copilot for ongoing cash flow monitoring.
Partner Workspace to serve the client portfolio; Cash Flow Diagnostic as the entry diagnostic; each company maintains its own subscription to the Copilot.
The two commercial entry paths are described in GTM · pp. 6, 8; per-company subscriptions are described in Pitch EN · p. 8.
The Copilot is the core product. The first documented implementation step is the Diagnostic in the Console, followed by reuse in the Partner Workspace, forecasts and alerts, financial questions on WhatsApp, and capture with approval. This sequence organizes implementation. Launch requires the complete planned scope to be delivered and validated; an individual increment does not replace product delivery.
Review screens and status · Review user journeys · Review open decisions
Delivery direction is established for the Diagnostic and its reuse in the Partner Workspace. The full inventory of screens, fields and layouts has not received blanket approval.
The Console diagnostic uses the selected company's context and is part of the planned scope. The local implementation documents work completed to date; its existence does not establish complete delivery.
Verify all planned report content, time horizons, organization by ERP app/company and communication of calculation limits. Validate the remaining work against the plan; the local 7/30/90-day implementation does not cap the scope.
The earlier inspection documented sign-in, the Hub, ERP integrations, ingestion progress and a reconciliation queue. E03 groups modal dialogs, rather than additional pages.
Use this foundation as implementation context. The inspection does not constitute redesign approval or demonstrate the complete financial operation.
The second increment reuses the Diagnostic for partners. Portfolio management and invitations are documented as existing code.
Specify the planned multi-company operation: metrics, columns, filters, prioritization, navigation and permissions. Portfolio management, access and diagnostic review must form the complete P01–P03 journey.
Forecasts and alerts precede financial questions. The materials specify daily monitoring over 7-, 30- and 90-day horizons, with WhatsApp as a channel.
Specify messages, horizon-specific rules, delivery frequency and any supporting Console interface. C01–C03 are touchpoints; they do not imply approval of three additional pages.
Document submission, draft creation, approval and ERP posting form the fifth increment. The PDFs require human approval and exception review.
Define the review interface or channel, permissions, correction and rejection behavior, and distinguish ERP draft persistence from effective posting. Tarik has not explicitly approved the technical write sequence.
The inspected R01 code opens a WhatsApp sales conversation. The materials describe the free diagnostic and the path to the Copilot.
Specify and validate the flow from diagnostic to offer, contract, automated billing and Copilot access. The existing sales contact does not replace this journey.
The implementation sequence is documented. All planned journeys are part of delivery; their detailed specifications and acceptance criteria still require definition and validation.
Sign-in → Hub → ERP connection → ingestion progress → Diagnostic is the documented local workflow.
To resolve: Implementation approval does not settle the commercial steps, report content or behavior when data is insufficient.
Portfolio → company → diagnostic is the direction for the second increment.
To resolve: Specify the liquidity view, navigation back to the portfolio and permissions required for the complete planned portfolio experience.
Daily monitoring over 7/30/90 days → contextual WhatsApp alert → recommendation and questions.
To resolve: Tarik approved these deliveries' position in the sequence; notification rules, availability and the final UI remain unresolved.
Content → draft pending approval → outcome and reconciliation is the documented commitment.
To resolve: The PDFs place drafts in the ERPs. Tarik has not approved a technical sequence for persistence, effective posting and approval.
The delivery commitment is established. The open items below specify how to implement and validate everything planned for all three products.
Launch criterion: complete planned scope, delivered and validated. The implementation sequence organizes the work; it does not authorize removing features, reducing report content or replacing the planned experience with a partial release.
Existing code is evidence of work completed. The product, screen and journey specifications define the delivery commitment. Implementation gaps remain work to complete.
Review product definitions and documented scopeSpecify the report that quantifies the gap between scheduled receivables and likely collections, making the cash timing mismatch clear.
Check fields and time horizons against the complete plan, then specify presentation, organization by ERP app/company and interpretation of the result. The local 7/30/90-day implementation is evidence to review; it does not define the product's scope ceiling.
Trace each requirement to the report and validate its data, calculations and J1 journey. A receivables analysis must not be presented as the company's full net cash position unless bank balances and accounts payable are included.
Specify behavior for each data state, retaining the planned experience and an accurate interpretation of the diagnostic.
Specify display criteria, messaging, the next action and recovery behavior for each state. Service availability, data coverage and statistical confidence each require distinct rules.
Validate every planned state with evidence. Missing data must not be treated as zero. Label partial results explicitly and do not present them as completion of the full journey.
Specify the complete path from free analysis to a recurring Copilot subscription.
Finalize screens, information, integrations and states for contracts, billing and activation. Confirm the deck's commercial wording against the current offer and specify WhatsApp's role in the journey.
Validate the journey from diagnostic to subscription and corresponding access. Assisted sales can support the customer, but must not replace the planned contracting and billing flow.
Specify the partner's portfolio experience and the transition to each company's diagnostic and financial workflows.
Specify metrics, columns, filters, comparison and prioritization rules, and permissions against the plan. P03's design still requires validation; this decision must deliver that experience in full, rather than reduce it to a shortcut to individual diagnostics.
Validate P01–P03 and J2 end to end, including navigation back to the portfolio and the boundaries of each role. The multi-company view must retain client data isolation; cash positions and ERP apps must not be aggregated without a defined rule.
Combine forecasts, alerts and conversation into an ongoing experience that explains issues and supports action.
Specify horizon-specific triggers, calculation of the receivables amount to bring forward, notification rules, message updates and the supporting Console interface included in the design. Resolve source discrepancies before finalizing these rules.
Validate C01–C03 and J3 using traceable data and rules, including answer availability and company context. A daily monitoring refresh does not, by itself, determine message frequency.
Specify the complete workflow from submitted content to draft, approval, outcome and reconciliation, with clear side effects and responsibilities.
Validate the distinction between draft records and effective postings in each ERP integration. Specify the review channel, permissions and ordering of draft persistence, approval and posting. The PDFs place drafts in the ERPs; the technical discrepancy must be resolved with evidence.
Validate C04–C06 and J4 end to end, including approval, correction, rejection and failure. The visible confirmation must match the actual effect in the ERP. Kaigoo does not transfer funds.
These six open items specify execution of the planned scope. None permits dropping features for implementation convenience. Completion requires evidence that the full planned journey works and has been validated; this document does not establish current production delivery.
The complete inventory, workflows and source audit are available below. Expand the relevant section or follow a reference link directly to the cited item.
Experience inventory by product. An item may be a page, modal dialog, message or interface responsibility; it does not automatically require a new screen.
Sign in, connect an ERP, track ingestion and select a company.
The user reported that onboarding had worked. The available evidence does not confirm the complete production journey or establish that the reconciliation queue constitutes the entire product.
Sign-in, email verification and registration completion, when required.
Enters an email address, follows the sign-in link and completes registration.
Access to the Hub and company context permitted for the account.
Earlier inspection described in the DOCX. COD · earlier code inspection
Companies, registration options and ERP connections.
Selects an existing company or starts a new connection.
The company's context or the start of onboarding.
Earlier inspection described in the DOCX. COD · earlier code inspection
Modal dialogs for companies and advisory firms, ERP connection validation and company identification.
Enters connection details and a name, then selects "Get started".
Data preparation begins. These are Hub modal dialogs, rather than additional pages.
Earlier inspection described in the DOCX. COD · earlier code inspection
Data preparation status within the Hub.
Tracks progress; the local increment provides a "View Diagnostic" action.
Data preparation visibility and access to R02 in the local implementation.
Earlier inspection described in the DOCX. COD · earlier code inspection
The day's queue items awaiting analysis.
Reviews items and performs the corresponding human action in the ERPs.
Progress in the financial workflow. This does not demonstrate autonomous execution or the complete draft lifecycle.
Earlier inspection described in the DOCX. COD · earlier code inspection
Documented routes: /entrar, /verificar-email, /completar-cadastro, /hub and /conciliacao. The /dev/progresso route is a development simulation and is excluded from the actual user journey.
Users review their company's diagnostic before deciding whether to subscribe to ongoing monitoring.
A sales call to action for the Diagnostic.
Selects the button on the marketing page.
Opens the sales WhatsApp conversation in the originally inspected version. This CTA did not demonstrate automated report delivery.
Earlier inspection described in the DOCX. COD · earlier code inspection · GTM · p. 5
Per ERP app: scheduled inflows, predicted collections and their difference over 7, 30 and 90 days, with the model date.
Selects "View Diagnostic" in the Hub and reviews the selected company.
Analysis of eligible receivable records. It does not cover the entire portfolio or full cash position; the actual query and complete journey still require validation.
The Console interface and 7/30/90-day horizons describe the local implementation; the PDFs do not prescribe this Diagnostic format. LOC · local implementation · GTM · p. 5
The next step after the free diagnostic.
Decides whether to subscribe to the Copilot through the commercial flow that is approved.
Transition to the recurring subscription, including the offer, contract, automated billing and confirmation. Interactions and integrations require specification and validation.
Move between the portfolio and individual companies, reusing the diagnostic and financial workflows.
The partner advisory firm's Hub and associated client companies.
Selects the company to monitor.
Enters the client's context within the user's access permissions.
Earlier inspection described in the DOCX. COD · earlier code inspection
Access management and invitation controls.
Reviews or manages existing relationships.
Access administration. An invitation does not, by itself, grant financial approval authority.
Earlier inspection described in the DOCX. COD · earlier code inspection · D04 · permissions
A multi-company view of diagnostics and attention signals.
Selects a company and opens its diagnostic or financial workflows.
Access to R02 and Copilot capabilities. Metrics, prioritization and navigation still require design.
The pitch describes free partner access and a separate subscription for each company. The Hub's presence in the code does not demonstrate the complete portfolio liquidity experience. PITCH · p. 8
Understand cash flow, ask financial questions and manage financial workflows with human oversight.
Daily monitoring over 7/30/90 days and a WhatsApp experience are already specified in the source materials. Layout, notification frequency, detailed criteria and the division between conversation and Console require definition.
Daily projections over 7-, 30- and 90-day horizons, highlighting cash pressure.
Reviews the forecast, the receivables amount to bring forward and the risk-based reference rate.
Context for a financial decision. The product does not provide credit, transfer funds or guarantee outcomes.
A projected cash shortfall, including its date, amount and context or cause, in the planned WhatsApp experience.
Reads the context and decides which issue to investigate.
A next action and suggested receivables amount to bring forward, where applicable. The conceptual trigger is a deficit or funding need; horizon-specific rules and notification frequency remain unresolved.
A WhatsApp conversation within the company's context.
Asks a question about the company's financial position.
An answer in clear Portuguese, the documented product language, explaining the issue, its cause and a suggested action. A question does not authorize a financial operation; simulations and operational limits require design.
A conversation for submitting an invoice, boleto payment slip, Pix payment receipt, photo or audio. Actual media support has not been validated.
Submits content from which a draft is prepared.
The PDFs describe drafts persisted in the ERPs. The earlier document placed approval before execution; this does not establish an explicit decision by Tarik on the technical write sequence.
The draft and supporting information required for review. The review interface or channel and user roles are not yet defined.
The manager reviews and approves; people resolve exceptions.
An entry subject to human approval. Distinguish draft creation, effective posting and funds movement before specifying the integration.
Operation outcome and daily reconciliation of bank statements, accounting entries and documents, including discrepancies and exceptions.
Checks the outcome and reviews exceptions such as interest, penalties and partial payments.
Continuation of the financial workflow and the data feeding forecasts and alerts. The integrated workflow has not been validated end to end.
The previously inspected WhatsApp handler recorded metadata and opt-outs. That inspection did not verify media processing, financial AI, drafts, approval or execution. The absence of evidence in that inspection also does not establish absence across the entire system.
Each step links to the inventory above. The diagrams describe the experience and its dependencies; they do not certify production functionality.
Goal: connect an ERP and obtain a useful view of likely collections.
This path combines the documented foundation with the local increment. R01's sales entry point opens WhatsApp; the connection from that conversation to the Console remains unresolved. The transition from R02 to R03 is a commercial decision, rather than an implemented subscription flow. D01 · D02
Goal: reach the correct company's context while preserving access boundaries.
P02 supports relationships and access. Viewing, editing and approval permissions require separate rules. P03 reuses the diagnostic and financial workflows, but its metrics and interactions have not been approved. D04
Goal: turn forecasts, alerts and conversation into an informed decision.
Forecasts, alerts and questions complement one another; they do not form a mandatory four-screen sequence. The product provides guidance on the amount needed and an indicative rate, without providing credit or transferring funds. PITCH · p. 4–6 · EST · p. 1
Goal: reduce manual work while retaining human review and exception handling.
The PDFs explicitly describe an ERP draft pending approval. The earlier document describes draft creation, approval and ERP recording, without establishing an explicit decision by Tarik on the technical write sequence. Draft persistence and effective posting must not be assumed to be the same operation. Resolve the distinction before implementing the integration. PITCH · p. 4 · GTM · p. 6–7 · D06
These labels describe the evidence behind this product map. They are not states that must appear in the customer interface.
An explicit commitment in the attached PDFs. It does not establish that the capability works in production.
Behavior identified in the earlier inspection and described in the DOCX. The code was not re-audited for this version.
A recent implementation retained locally. Technical checks do not establish validation of the actual user journey.
A proposed structure or behavior for discussion. It does not constitute approval of an additional screen.
A decision requiring product, business or engineering validation before it can be finalized.
Implementation context and user direction given after the PDFs. Conflicts are surfaced rather than removed.
This review read the three PDFs directly. Code and local implementation evidence comes from the earlier document, with its limitations retained.
A direct reading of the three PDFs compared with the earlier document. The corrections below distinguish commercial commitments, technical evidence and subsequent decisions.
The Copilot is the core product; the Workspace serves the partner channel; the Diagnostic is the acquisition product. The Forecasting Engine is a shared capability, and the five increments define the implementation sequence.
Readiness, triggers, statistical weights and ERP draft timing contain explicit inconsistencies. None has been silently treated as an implemented requirement.
Earlier documentThe third source was missing from the reference set, leaving part of the intelligence capability undocumented.
This versionThe shared engine and its limits are included in the architecture. Complete liquidity monitoring includes current balance + expected collections − accounts payable, rather than predicted inflows alone.
Earlier documentC01, C02 and D05 previously left the channel and time horizons entirely open.
This versionThe sources specify daily monitoring over 7/30/90 days and WhatsApp. A supporting UI, notification cadence, precise triggers and horizon-specific rules remain under discussion.
Earlier documentAlerts were described only as signals to investigate.
This versionThe specification now includes the issue, cause, shortfall date and amount, next action, suggested receivables amount to bring forward and a reference rate where applicable. This does not imply credit provision or funds movement.
Earlier documentChannel and format were left entirely open, while the local implementation used the Console and 7/30/90-day horizons.
This versionThe GTM plan specifies a free automated report. It does not fix detailed time horizons or the delivery channel; the Console and per-app 7/30/90-day calculations remain identified as local implementation choices.
Earlier documentThe main product map omitted pricing, free partner access and data isolation.
This versionThe pitch describes a Copilot subscription at BRL 399 per company per month, with no long-term commitment or implementation fee; a free Partner Workspace; separate company subscriptions; and data isolation by CNPJ. These are source terms, rather than evidence of current commercial availability.
Earlier documentC04/C05/J4 previously placed approval before any ERP operation.
This versionThe PDFs place drafts in the ERPs before they take effect, subject to manager approval. The earlier document placed approval before execution. That wording does not establish an explicit decision by Tarik; persistence, effective posting and approval must be reconciled.
Earlier documentQuestions and outcomes were described generically; the existing queue risked standing in for the full commitment.
This versionAnswers must explain the issue, cause and action. Daily reconciliation matches bank statements, accounting entries and documents, with human exception review. The current queue does not establish delivery of that commitment.
Earlier documentThe commercial structure and implemented screens were presented without a clear distinction.
This versionThe GTM plan recommends separate entry paths for owners/managers and accountants/advisors. The Hub, email sign-in, modal dialogs, invitations and routes come from the earlier code inspection, rather than prescriptions in the PDFs.
Earlier documentD03 previously left all data handling unresolved.
This versionThe sources document a blend of payer-specific and industry data, a fallback at n=0 and sample-based confidence. These rules do not resolve missing balances, accounts payable or ingestion, or validate the complete diagnostic.
Earlier documentThe DOCX described engineering as paused pending review and cited local tests as evidence.
This versionThe latest direction resumed local engineering in parallel. The blueprint remains under review; tests, marketing claims and the existence of code do not establish production delivery.
These issues remain visible for review. The audit does not choose a technical interpretation without additional evidence.
The report describes available analytical assets, but also lists alert activation, rate guidance and collections as next steps. The English pitch describes the product workflow and states that the engine, pipeline and WhatsApp channel are in production; contracts, automated billing and monitoring appear as work still to complete. Finding: the commitment is documented; this review does not demonstrate complete current delivery to customers.
The report references 7/30/90 days, but explicitly defines valor_necessidade_caixa = abs(caixa_projetado_90d) and mentions 30/90 days in its next steps. It does not define a consistent contract for each horizon. To specify: the field, sign convention, amount calculation and deduplication rule.
The pitch illustrates a BRL 18,400 deficit and a suggestion to bring forward BRL 12,000 in receivables, without explaining how one amount produces the other. It explicitly identifies names and amounts as fictional. Do not infer a formula, calibration method or risk rule from this example.
The PDFs place drafts in the ERPs; the earlier document places ERP recording after approval. Neither source verifies the API mechanism. Retain human approval and specify the sequence before implementing writes to an actual ERP.
The Forecasting Engine report and pitch use different volumes, entities and portfolios. The pitch includes a September 2, 2026 snapshot; the GTM document also abbreviates "payers and suppliers". Do not combine these figures into growth claims, current totals or customer metrics for this site.
The report classifies missing history as critical under its commercial policy while also describing an industry-based statistical fallback. These may be distinct policies. Reconcile confidence, risk and user messaging; missing history does not establish default.
The presentation claims 90% payer-specific history above 30 purchases and 100% personalization. However, n/(n+12) yields 71.43% at n=30 and reaches 90% only at n=108. Explicit inconsistency: do not repeat those percentages as mathematical facts.
Accounting conservation, distribution normalization and a claimed absence of temporal leakage do not demonstrate accuracy or calibration on future data. Describe forecasts and estimates, without promising exact collections, no false alerts or guaranteed outcomes.
The sources differ on acquisition cost ranges and baseline analyst capacity, while retaining a target of up to 40 companies. Do not present them as measured CAC reductions, verified productivity or contractual limits of the Partner Workspace.
The report describes a risk score and a preventive collections sequence; collections deployment also appears as a next step. Do not recast these as a fourth product, a mandatory additional screen or a newly approved increment.
References use physical PDF page numbers. The audit checked the attached documents; it did not revalidate repositories, databases, metrics or the production product.
12 pages; content on pp. 1–8. Pages 9–12 were checked and are blank.
pp. 1–5: roles of the three products · p. 5: free Diagnostic and automated report · p. 6: two commercial entry paths · pp. 6–7: capture, drafts and reconciliation.
16 pages. The English version supplied for this review was checked on October 2, 2026. It replaces the earlier version's Portuguese pitch reference. Commercial terms and metrics belong to the deck's dated snapshot.
pp. 4–6: capture, reconciliation, forecasting, human approval and conversation · p. 8: BRL 399 per company per month, free analysis and free Partner Workspace · pp. 9, 11: channel and commercial development · pp. 13–16: historical data and glossary.
4 pages. The formula, table and diagram on pp. 2–3 were checked visually.
p. 1: liquidity calculation and recommendations · pp. 2–3: risk, confidence, payer/industry blending and weights · p. 4: next steps. Inconsistencies between commitments, status and formulas remain documented in V01–V10.
The earlier eight-page document audited here. It is retained as a reference; this version is the requested website.
IDs E01–E05, R01–R03, P01–P03, C01–C06 and J1–J4 are retained for traceability. Code and local work described in that document are prior evidence, rather than a new inspection.
An earlier inspection, not repeated in this audit. The existence of a component does not establish end-to-end delivery.
Console: CtaSection.tsx, onboarding/empresa/route.ts, OnboardingModal.tsx, hub/page.tsx, progresso-store.ts, fila-do-dia.tsx and whatsapp-inbound.ts. Intelligence layer: forecasting models identified in the earlier inspection.
An earlier record, subject to engineering resumed in parallel. This review does not verify a query against actual data, production deployment or the complete journey.
EmpresaBloco.tsx · RaioX.tsx · API /api/raio-x · read access to estimador.p1_gatilho_diario · branch codex/raio-x-v1. Documented checks: 1,522 tests, TypeScript, Biome and build. These checks do not establish access to actual customer data.
Tarik approved the five-increment sequence and the resumption of local engineering in parallel. Initial work on the Diagnostic in the Console was authorized; this does not constitute blanket approval of layouts, fields or time horizons.
The three-product structure comes from the PDFs. Interface details and the technical ERP write sequence have not received blanket approval from Tarik. Source discrepancies do not authorize product deployment, funds movement or additional permissions.
Direction given on October 2, 2026: deliver the complete planned scope. Open decisions specify how to implement and validate it; they do not authorize feature cuts, feature deferrals or substitution with a partial release. Launch requires completion of the full plan. View scope direction and execution specifications.
Product terminology: specific vendors named in the materials are represented by the category ERPs. The audiences are companies and advisory firms. Technical coverage depends on the integrations actually delivered.
This site is a product alignment document. Publishing it does not deploy the Kaigoo product or enable financial operations. Actual customer data was not used as a functional demonstration.
Demonstration data supports technical validation and is not a manual step required of customers. There is no evidence that customers must prepare synthetic data to use the journey.
Repositories cited in the earlier document: techteu-console and camada-inteligencia-kaigoo. Access depends on the account's permissions.