Milestone and Use-of-Funds Monitoring: Governance for Financial Institutions

What investors need to see between approval and results

A useful milestone combines an observable deliverable, acceptance criteria, a designated verifier and a contractual consequence. “Expansion underway” reveals little. “Equipment delivered, identified and installed at the agreed location, with a documented operating test” describes something that can be checked. Even then, physical delivery alone does not establish customer demand or future cash generation. Monitoring must distinguish physical execution, financial compliance and economic viability. Evidence supporting one part of a project should not become an unsupported conclusion about the entire investment.

Every number in this article is hypothetical and educational. None represents a market rate, a KEZS transaction or an observed result. The discussion concerns an appropriately authorized institution assessing origination channels, possible correspondent arrangements and business financing products. Investors conduct their own diligence and decide within their mandates. Organized documentation can reduce information gaps, but it cannot eliminate fraud, delays, losses or disagreements over interpretation. The purpose is to support decisions under uncertainty, including decisions to investigate, change conditions or stop further funding.

Define the financial instrument before choosing milestones

The same project can receive funding through instruments that create different rights. For a loan, monitoring relates to repayment obligations and agreed restrictions on use. For a receivables purchase, it concerns whether the claims exist, meet eligibility requirements and support the purchaser's legal position, alongside payment behavior. Equity exposes investors to business risk and makes information, voting and exit rights relevant. A convertible instrument adds conditions for a possible conversion into equity. Controls suitable for one instrument should not automatically be transferred to another.

A sale the business hopes to make is not an existing, eligible receivable. An order may be canceled, depend on delivery or contain unmet conditions. Reports should distinguish an expression of interest, a contract, delivery, invoicing and payment. Legal structure also matters in Brazil: a MEI, or individual microentrepreneur, cannot admit equity partners while remaining a MEI. A donation to a nonprofit is neither an equity investment nor a promise of financial return. These distinctions change what milestones mean, what evidence is relevant and what rights the funder actually acquires.

KEZS connects entrepreneurial projects and potential funders through technology and intermediation; it does not assign a credit score. A project-based method can organize purposes, stages and documents for assessment without replacing the institution's judgment. Public securities distributions and regulated financial activities require appropriate structures and providers. Technology, a commercial presentation or a CNAE economic activity classification does not authorize those activities. Explaining a monitoring method therefore does not establish that a platform has regulatory permission, banking partnerships or any particular operating arrangement.

Institutional policy, decision authority and retained responsibilities

An institution should not rely on an originator's commercial policy to determine what it finances. Its own policy should define permitted purposes, eligibility criteria, concentration limits, minimum documentation, disbursement conditions and review triggers. It should identify which decisions may be automated, which require an analyst and which need a higher approval level. Investors need evidence that these rules operate in practice, including in approved transactions and exceptions. A policy manual alone does not show whether staff apply its requirements consistently or record departures from them.

A correspondent arrangement depends on a contract and an agreed scope with the institution. The agreement should specify permitted activities, channels, operational duties, data access, remuneration and supervision. Receiving an application is not approving credit; collecting documents is not completing diligence. The financing institution retains the responsibilities assigned to it under applicable regulation. Delegating collection, product presentation or execution monitoring must preserve those responsibilities. Investors should be able to distinguish who performs a task from who remains accountable for the associated decision and control.

In practice, commercial staff may request a change, an analyst may assess its effects and an authorized decision-maker may approve it. Someone paid for origination should not certify, without independent control, the milestone that releases their own commission. Exceptions need reasons, an expiry or review point and an identified approver. Frequent recurring exceptions may indicate that policy no longer reflects the risk actually being accepted, even if each individual case appears defensible. Monitoring should reveal that pattern rather than treating every exception as unrelated.

Turn the project plan into an evidence matrix

Start by breaking the financed purpose into identifiable uses. For each use, an evidence matrix records the budget, approved supplier or supplier category, relative deadline, funding source and expected documentation. It then connects each disbursement to a release condition and subsequent verification. This avoids requesting proof only after money has left, when the contract never clearly established what would count as sufficient evidence. The matrix also helps identify an unfunded step or a verification requirement that cannot realistically be met at the proposed point in execution.

Each record should answer the right question. A quotation shows an offered price; a Brazilian fiscal invoice, or nota fiscal, records a declared transaction; a bank receipt shows a transfer; technical acceptance addresses delivery or performance. None automatically substitutes for the others. Significant equipment purchases may require matching asset identification, fiscal documentation, the payment beneficiary and receipt confirmation. For services, a generic progress report may be inadequate unless it identifies deliverables, responsible parties and acceptance criteria. Document volume is less useful than a clear connection between evidence and the claim being verified.

Verification costs should reflect materiality and risk. Repeating an expensive inspection for every small purchase in a dispersed budget may be disproportionate. Conversely, superficial sampling may be inadequate for an asset consuming most of the financing. The matrix should explain when full checking, sampling or additional diligence applies. Investors gain useful information when reports state what was checked, how it was checked and which gaps remain. A simple “completed” label conceals those distinctions and makes it harder to assess the strength and limitations of the conclusion.

KYC and financial crime controls throughout execution

Knowing the customer is not limited to opening the relationship. Institutions must apply identification, customer assessment and financial crime prevention procedures appropriate to the product and their regulatory framework. Changes in representatives, ownership, payment beneficiaries or economic activity may require updates during execution. The objective is not to treat every change as misconduct. It is to recognize when reality no longer matches the information used in the original decision and determine whether the difference affects eligibility, payment controls or the assessment of risk.

A supplier replacement illustrates the issue. A new provider may offer a better price and timetable, yet may also have an undisclosed connection to the borrower or lack apparent capacity to perform the service. Before payment, the process should support checks on identity, the commercial relationship, the reason for substitution and consistency with the financed purpose. Third-party accounts, unusual payment splitting and requests to return funds through a different route warrant assessment under institutional policy. These signals call for investigation rather than premature conclusions about wrongdoing.

Within its contracted scope, an originator can provide information and flag inconsistencies. Assessment and any required action remain with the competent responsible parties. Suspicions should not circulate indiscriminately through commercial reports, and data handling should respect purpose, necessity and access controls. Investors need to know whether relevant changes are detected and whether their treatment is traceable. Unrestricted copies of personal documents are not necessary to understand the strength of the process. A controlled account of the checks and resulting decisions can be more informative.

A complete example: budget, funding sources and three disbursements

Consider a hypothetical business expanding a production unit with a R$600,000 budget. Planned uses are R$300,000 for equipment, R$120,000 for installation, R$100,000 for initial inventory and R$80,000 for a cash reserve. Funding consists of R$180,000 from the business's own resources and a R$420,000 loan. Uses and sources both total R$600,000. That balance is necessary to complete the funding plan, but it does not establish that prices are reasonable, the reserve is sufficient or the business can support the resulting debt.

The business's contribution covers R$120,000 of the equipment and R$60,000 of the reserve. The loan has three disbursements: R$180,000 completes the equipment purchase; R$140,000 pays R$120,000 for installation and R$20,000 for inventory; and R$100,000 purchases the remaining R$80,000 of inventory and adds R$20,000 to the reserve. Equipment therefore totals R$300,000, inventory R$100,000 and the reserve R$80,000, with no expense funded twice. The institution should be able to reconstruct this allocation from the contract and records without relying on a later verbal explanation.

The first release requires proof of the business's contribution and the purchase documentation conditions. The second requires equipment delivery and validation of the installation plan. The third requires acceptance of installation and satisfaction of the inventory purchase conditions. These are hypothetical contractual choices for an educational example, not mandatory rules. Each institution must assess whether the sequence fits the instrument, the borrower's capacity and the risk of leaving stages only partly funded. A balanced budget can still contain a funding sequence that is impossible to execute.

The timeline reveals risks hidden by the budget

In the initial week, the business contributes R$180,000 and pays a R$120,000 equipment deposit, leaving R$60,000 in cash. The first R$180,000 loan disbursement arrives and pays the equipment balance, preserving R$60,000. In week four, the second disbursement adds R$140,000; installation payments of R$120,000 and inventory payments of R$20,000 again leave R$60,000. In week eight, R$100,000 arrives and R$80,000 goes toward inventory. The final balance reaches R$80,000. Each movement follows the allocation established in the funding plan.

This timeline includes only the movements in the project budget. Ordinary operating expenses, taxes, financing charges and any revenue require separate forecasting and integration into the business's overall cash position. Treating this project balance as unrestricted money for any obligation would be misleading. The institution must also verify that the R$60,000 retained during implementation actually exists and is not committed elsewhere. Reports should distinguish the bank balance, amounts reserved for a particular purpose and funds genuinely available for use. The same cash cannot safely support several commitments at once.

Coordination creates another risk. If the installer requires payment before equipment delivery, the second disbursement condition could prevent the schedule from proceeding. That problem should be addressed before contracting. Depending on the institution's assessment, the response might change supplier terms, adjust a release condition or require another funding source. An excessively rigid milestone can stall a viable project; an overly permissive one can advance money without enough protection. The contract should make this trade-off explicit, including who can approve a change and what evidence is needed.

Separate execution, performance and repayment capacity

Delivered equipment and purchased inventory demonstrate partial execution, not efficient use, positive margins or future repayment. After installation, monitoring may examine actual production, waste, completed sales, collection periods and recurring expenses. Indicators should connect to the assumptions supporting the financing decision. If the business case depends on reducing waste, revenue alone may hide a deteriorating margin. If it depends on faster delivery, installed capacity says little without evidence of actual turnaround. Monitoring becomes useful when it tests the mechanism expected to generate cash rather than merely counting completed purchases.

For a hypothetical debt-service calculation, assume the fully disbursed R$420,000 is repaid through twelve monthly principal installments of R$35,000, with interest of 1% per month on the preceding outstanding balance. The first payment is R$39,200: R$35,000 principal plus R$4,200 interest. The final payment is R$35,350. Interest across this repayment phase totals R$27,300, making total payments R$447,300. These figures cover only the stated amortization phase. Implementation-period charges, taxes and other costs are excluded and would need to be added according to the contract. The rate is an arithmetic assumption, not a market quotation.

If projected monthly cash available before servicing this debt is R$50,000, the initial cushion is R$10,800. If available cash falls to R$30,000, the first payment has a R$9,200 shortfall, even if every physical milestone is complete. These are scenarios, not results. The R$80,000 reserve may absorb part of a timing mismatch, but consuming it does not make a persistently cash-deficient operation sustainable. Investors should examine where cash comes from and whether it recurs, rather than treating a temporary balance as evidence of lasting repayment capacity.

Settlement and reconciliation: establish the money trail

Disbursement authorization, financial settlement and reconciliation are distinct events. A tranche may be approved without being transferred. A submitted payment instruction may fail or be returned. A bank debit may lack the correct matching entry in the contract records. The institution needs to identify the evidence for each state and assign responsibility for resolving differences. Reports should avoid describing money as “disbursed” when it has only received internal authorization, especially if the borrower has already entered into supplier commitments based on the expected funding.

In the example, the second tranche requires tracing an incoming R$140,000 and two uses: R$120,000 for installation and R$20,000 for inventory. Each movement should connect to the transaction, tranche and supporting document. Direct supplier payments require beneficiary and account checks under the applicable procedure. If funds pass through the business's account, the institution must assess how far traceability actually extends. A partial statement may not show commingled money, intermediate transfers or later refunds. Evidence should support the entire conclusion being reported, rather than just its first step.

Reconciliation also covers reversals, refunds, discounts and value differences. If a supplier refunds R$10,000, records cannot continue showing the original use as fully substantiated. The refund must be recorded and an authorized decision made, under the contract, about reapplication, retention or exposure reduction. A dedicated account may simplify monitoring, but it does not by itself establish legal segregation or protection against creditors. Banking controls and contractual rights require separate assessment. A clear money trail helps explain what happened; it does not automatically create an enforceable right to recover the funds.

Proportionate alerts and a routine that leads to decisions

A useful dashboard organizes exceptions and actions. It should distinguish a delivery still within its deadline, a justified delay that has been accepted, rejected evidence and a blocked disbursement. Colors without stable definitions create an appearance of control while undermining comparisons. Investors need to know how long an issue has remained open, how much exposure it affects and what decision is expected. Update frequency should follow the project's cycle rather than the convenience of monthly reporting. A critical change may need attention before the next reporting package is prepared.

Triggers can combine elapsed time, financial amount and changed assumptions. A minor delay outside the critical path may require only an update. A delay preventing equipment installation before a payment obligation may require a revised cash forecast. Budget overruns consuming the reserve, significant supplier changes or persistently weaker collections call for specific analysis. Numerical alert thresholds are institutional policy choices requiring calibration; they are not universal requirements. An alert should identify why the change matters to execution or repayment, rather than merely indicate that a number crossed a line.

Every alert needs an owner, a treatment deadline and a closure record. Suspending a tranche, accepting additional evidence and revising a schedule are different decisions with different consequences. Monitoring records should preserve the evidence available when the decision was made, including when the eventual outcome was poor. Judging solely by the outcome encourages hindsight. Process quality depends on decisions consistent with the facts then known and on recognizing promptly when an assumption no longer holds. Recording a decision's reasoning also allows independent reviewers to distinguish an informed risk acceptance from an overlooked problem.

Adverse scenarios and the limits of staged funding

Return to the R$600,000 budget. Suppose installation rises from R$120,000 to R$150,000, with every other use unchanged. The project now requires R$630,000, creating a R$30,000 funding gap. An additional R$30,000 contribution raises the business's own funding to R$210,000 while the loan remains R$420,000, preserving the planned R$80,000 reserve. Alternatively, paying the overrun from the reserve reduces it to R$50,000, requiring a fresh liquidity assessment. Both approaches can balance the revised plan, but they leave different protection against future cash shortfalls.

Recording “budget adjusted” is insufficient. The institution should investigate whether the increase reflects an initial estimate error, a newly discovered technical condition, weak negotiation or changed scope, and whether other costs remain omitted. A promised contribution is not available money. Without a verified source to close the gap, the next disbursement may increase exposure without securing completion. Yet an abrupt stop may leave equipment uninstalled and reduce recoverable value. The decision requires comparison of concrete alternatives, including their funding needs, likely delays, costs and consequences if completion still fails.

Other adverse scenarios include unsuitable equipment, an insolvent supplier, inventory losing usefulness and slower sales. Staged releases limit exposure before certain deliverables but do not guarantee recovery of money already paid. Equipment resale proceeds may fall below its purchase cost, and removal can generate further expenses. Guarantees, insurance and third-party commitments help only within their validity, coverage and practical enforceability. Investors should examine possible losses and exit costs alongside continuation conditions. A control that improves visibility may still be unable to prevent a commercial failure or make an illiquid asset readily saleable.

Disputes, renegotiation and change control

Entrepreneur, originator and funder may disagree about whether a milestone has been met. A supplier calls installation complete; the business identifies defects; the analyst requests a test required by the contract. Sound handling starts by separating facts, interpretation and economic interests. Photographs may show that equipment is present without resolving a technical dispute. An independent assessment may be more useful than further documents produced by the same parties, provided its scope and cost are proportionate. The question is which evidence can resolve the disputed acceptance condition.

The contract should provide a route for challenging conclusions and supplementing evidence. During review, participants need to know which payments may continue, which remain suspended and who has authority to decide. Poorly communicated blocks can worsen delays and create avoidable costs. Informal releases weaken agreed conditions. Recording the dispute allows review without turning monitoring into an argument over recollections of meetings. Communications should identify the unresolved issue and required next step clearly enough for affected parties to act, while preserving appropriate confidentiality and the institution's decision authority.

Renegotiation replaces part of the agreement with newly approved terms; it does not erase earlier noncompliance. If the schedule changes, retain the original version with the justification, cash-flow effect and authorized decision. Options may include additional capital, reduced scope, a different payment sequence or reassessing continuation. None is neutral: more time can increase exposure, reduced scope can lower expected revenue and additional requirements can make execution unworkable. Reports should explain those effects. Changing the written plan is meaningful only if the revised funding, timing and operational assumptions are credible.

Assess the origination channel and its conflicts of interest

Selecting an origination channel requires looking beyond application volume. Document quality, information accuracy, rework and post-contract performance matter. Comparisons must account for differences in customers, products and portfolio maturity. A channel serving more complex businesses may generate more outstanding issues without necessarily performing its role worse. Likewise, low delinquency in a very young portfolio does not establish sustained quality. Investors need comparisons that recognize how long transactions have been exposed to repayment risk and whether the channels are performing equivalent tasks.

Remuneration influences behavior. Commission concentrated at signing may encourage volume and weaken attention to later execution. Payment tied to a milestone can create pressure for premature acceptance. Controls require defined responsibilities, quality criteria and independent supervision, with clear rules for inaccurate information and process failures. Total channel cost includes diligence, integration, support, verification and exception handling. A seemingly low commercial fee can conceal substantial institutional operating costs. The assessment should therefore consider both the incentive created by the fee and the work required to use the channel responsibly.

Investors can review comparable transaction groups for documentation time, suspended releases, budget changes, losses and recoveries, with clear denominators and observation periods. They should also ask whether rejected or withdrawn applications disappear from statistics. Project-based intermediation may make information more organized and comparable, but its usefulness needs support from actual processes and available records. A platform's existence does not establish a banking partnership, a correspondent contract or operating capacity. Proposed methods and documented performance must remain distinguishable, and apparent efficiency should be assessed alongside unresolved issues and the cost of controls.

Transparent documentation with appropriate data protection

Transparency does not mean making every document available to every participant. Investors need enough information to understand exposure, progress and limitations; personal documents, banking data and sensitive commercial information require protection. Access design should distinguish collection, analysis, approval and consultation of consolidated results. Sharing purposes must be clear, as must retention, correction and disposal procedures under applicable obligations. A monitoring arrangement should explain who needs access to which evidence, rather than assuming that broader distribution automatically improves oversight or satisfies diligence requirements.

A report can show tranche amount, use category, reconciliation status and verification conclusions without unnecessary personal data. When diligence requires an original document, access can be controlled and logged. Identify versions so an old proposal cannot be mistaken for the current contract. Preserve provenance as well: a borrower-supplied file, third-party confirmation and independent verification carry different evidential weight. Losing that distinction can turn a declaration into an apparent verified fact. Reports should make clear both the underlying source and the extent of the checking actually performed.

Cryptographic records can help detect subsequent changes to files, but integrity is not truth. Blockchain does not establish delivery, legal ownership, reserve solvency or a payment guarantee. Private documents should not be placed on public blockchains. Before adopting a technical solution, the institution should define the problem it addresses, who may access data and how errors will be handled. A technically immutable record of incorrect information still needs correction and context beyond an integrity check. Technology should serve a defined control need, with its cost and limitations assessed alongside its potential usefulness.

Correct operational errors without destroying the audit trail

A common mistake is linking the same fiscal invoice to two tranches. In the example, the R$120,000 installation invoice could appear as support for both the second and third releases, falsely suggesting adequate documentary coverage. Correction begins by identifying duplication and recalculating the uses actually substantiated. If the third tranche has not been paid, the issue may block release under the contract. If it has, the institution must locate the actual use, assess any breach and refer the decision to the proper authority. Reusing evidence does not substantiate another expenditure.

Another error is classifying a R$20,000 inventory transfer as cash reserve. Financial totals may still balance, but liquidity information will be wrong. Correct the category, identify supporting documentation and update affected reports. Changing the displayed number is insufficient: determine whether earlier decisions relied on the error. If they did, reassess those decisions and communicate the impact to relevant recipients while preserving history. This matters because apparently available cash may already have been spent, changing the assessment of whether the business can meet its next obligation.

The audit trail should record who changed the data, why, the supporting evidence and the effect. Its purpose is to reconstruct the transaction without preserving an error as truth or erasing its occurrence. Second-person review may be appropriate for changes affecting exposure, milestone compliance or available cash. Quality reports should distinguish missing documentation, operational mistakes and indications of intentional misconduct. Combining these categories can exaggerate a correctable issue or minimize a serious irregularity. The response should reflect what is known, what remains uncertain and which decisions the correction affects.

A practical basis for the investor's decision

Before accepting that a portfolio is well monitored, an investor can request reconstruction of a transaction: approved purpose, sources, uses, milestones, releases, payments, reconciliation and exception decisions. Do the records tell a coherent story without improvised adjustments? Then examine cases involving delay or budget changes, where process limitations become visible. Sampling only transactions without outstanding issues reveals little about the ability to handle difficulties. The reconstruction should connect evidence to decisions, including decisions made when the available information was incomplete or the interests of participants differed.

Project closure also needs criteria. Completing installation does not necessarily end credit monitoring; repaying debt does not establish that business objectives were achieved. A final report should identify substantiated uses, differences addressed, remaining available funds and continuing obligations. Lessons may inform future policy but must remain separate from changes to an existing contract. Revising an assumption should produce an identifiable action, not merely new report wording. The institution should be clear about what has ended and what still requires attention, so completion labels do not conceal unresolved obligations.

Reference entry points include Banco Central do Brasil guidance on correspondents, CVM Resolution 88 on investment crowdfunding and CVM Resolution 175 on funds, as relevant to the instrument. Portal do Empreendedor and ANPD provide starting points for MEI and data protection questions. These sources are not proof of KEZS authorization or substitutes for case-specific analysis. The practical questions remain: what right is acquired, who decides, what evidence supports each release and what happens if the plan fails? Monitoring supports decisions under uncertainty. It provides no guarantee of approval, repayment, profitability, liquidity or preservation of capital.

Frequently asked questions.

References and context

Institutional reference points for checking applicable rules and guidance. Calculations and scenarios are original examples, not quotes, historical results or individualized recommendations.

• Banco Central — correspondentes no País ↗
• CVM — Resolução 88, texto e alterações ↗
• CVM — Resolução 175, fundos de investimento ↗
• Portal do Empreendedor — perguntas sobre MEI ↗
• ANPD — orientações e perguntas frequentes ↗

AI-assisted educational content. Decisions require verification of the information, instruments and providers in each actual case.