Programs combining business preparation, product development, commercial support and possible investment offer an appealing proposition: bringing together capabilities that a young company might struggle to hire and coordinate. For investors, however, this concentration raises a question before any return projection: what retains value if the program's main promises fail to materialize? Answering it requires separating the supported business, the services purchased and the financial instrument used.
An investment must withstand a review of its assumptions
Downside scenario analysis examines plausible combinations of unfavorable events. Sensitivity analysis measures how much a conclusion changes when one assumption changes. The approaches complement each other. Reducing revenue alone reveals its influence on cash; combining a product delay, weaker sales conversion and the need to redo work exposes a mechanism of loss that an isolated change can conceal.
The objective is not to produce a pessimistic forecast with an appearance of precision. It is to identify the conditions supporting an investment decision, the signals that would justify revisiting it and the rights that would help during that review. A company can remain technically solvent while becoming an unacceptable investment. Conversely, a manageable delay need not destroy the investment thesis.
All figures in this article belong to an entirely hypothetical example. They are neither market rates nor observed results, and they do not constitute an investment recommendation. The focus is how an investor evaluates an integrated program: recognizing dependencies, distinguishing services from capital and negotiating consequences before a shortage of money narrows the available choices. Cash, evidence, contracts and the ability to correct problems must be assessed together.
Unpack the arrangement to understand rights and payments
A useful starting document maps the economic relationships: who receives the investment, who provides each service, who pays, who holds equity and who can approve contracts between those parties. One group might operate the accelerator, control a development company and manage an investment vehicle. That arrangement does not establish wrongdoing, but the parties cannot be treated as one counterparty with a single interest.
Equity, convertible instruments and debt require their own terms. Equity involves the rights and risks of ownership. A convertible depends on its conversion conditions, triggering events and other agreed provisions. Debt creates payment obligations under its governing instrument. A service contract must separately define scope, compensation and responsibility for execution. Signing one document does not automatically settle the rights and obligations covered by the others.
Providing services does not automatically constitute an investment of capital. When a program advertises a value for its support, investors should establish how much enters the company as cash, how much represents work, how much is a discount and how much depends on future purchases. Available consulting hours, conditional credits and spendable cash have different effects on operating continuity.
Payment obligations must also be distinguished from execution risk. A provider may be entitled to payment for accepted work even if the company cannot sell its product. Conversely, a commitment to commercial support may contain no obligation to obtain customers. Downside analysis starts with these contractual differences because they determine who bears the cost when product delivery, sales and financing stop progressing together.
Build an assumption register that can be challenged
A projection becomes reviewable when its assumptions have a source, an owner and a revision condition. For each material assumption, record the value used, supporting evidence, plausible range and consequences of being wrong. A claim that commercial support will shorten the sales cycle should identify the reference cycle, customer profile and specific activity expected to produce that improvement.
Classify information by what it establishes. A reconciled payment is an observed fact; an advanced negotiation is a sales opportunity; an expected renewal is an assumption. A signed contract may support a projection, but its contents still require examination. Acceptance conditions, cancellation rights, payment timing and implementation obligations can change the economic value attributed to the agreement.
The register should make dependencies explicit. Projected revenue may require an unfinished feature, which requires access to customer data, which requires a contractual review. Treating these stages as independent entries creates false diversification. If one unresolved issue blocks four customers, the company has four commercial exposures connected to a common cause. Progress on unrelated tasks does not remove that shared bottleneck.
A productive investment discussion allows assumptions to be challenged without turning disagreement into a personal judgment. A commercial manager may defend a higher conversion estimate while finance uses a lower figure for cash planning. Both can coexist if their purposes are clear. The mistake is presenting the commercial expectation as fact while confining uncertainty to an inconspicuous presentation note.
The numerical example: separate investment, services and budget
Consider a company receiving R$ 600,000 from an investor for a 20% ownership interest immediately after the investment. In this simplified example, with no other instruments, options or adjustments, the implied valuation is R$ 2.4 million before investment and R$ 3 million afterward. The operation starts with R$ 600,000 in cash and no other available balance. Ownership guarantees neither capital recovery nor distributions.
Separately, the company purchases six months of program services for R$ 180,000, payable in six monthly installments of R$ 30,000. Total monthly cash outflows are budgeted at R$ 120,000: R$ 90,000 for staff and other activities, plus R$ 30,000 for this contract. From month seven onward, the model reserves the same R$ 30,000 monthly for continued or replacement support. Projected spending therefore does not automatically fall when the initial contract ends.
Operating cash receipts are assumed to be R$ 40,000 per month. This is an assumption about money received for scenario calculations, rather than a label for hoped-for sales. Net monthly cash burn is therefore R$ 80,000: R$ 120,000 of outflows minus R$ 40,000 of inflows. The R$ 180,000 service bill is already included in those outflows and must not be added again as a separate funding requirement.
The investor contributes R$ 600,000, and the provider will receive R$ 180,000 if its installments become due and are paid. Even if both belong to the same group, the company does not have R$ 780,000 of cash available. Nor does the service price prove that the work adds R$ 180,000 to business value. Keeping these dimensions separate prevents double counting and allows the investment and purchasing decisions to be assessed individually.
The cash calendar reveals when decisions must be made
In the base case, cash ends month one at R$ 520,000, month two at R$ 440,000 and month three at R$ 360,000. It falls to R$ 280,000 after month four, R$ 200,000 after month five, R$ 120,000 after month six and R$ 40,000 after month seven. At a constant R$ 80,000 monthly net burn, cash runs out halfway through month eight. The calculated runway is 7.5 months, not twelve.
This calculation assumes receipts and payments occur evenly within each month. That simplification makes the arithmetic readable. In practice, payroll due early and customer receipts arriving late can cause a cash shortfall before the average suggests. Investors should request a more detailed payment calendar when the gap between minimum cash and mandatory payments becomes narrow. A positive month-end projection does not establish that every earlier payment can be made.
The model already identifies a decision deadline: management cannot wait until cash is exhausted to discover whether the next stage can be financed. Suppose it adopts a hypothetical minimum operating reserve of R$ 160,000, equivalent to two months of base-case net burn. Cash crosses that threshold during month six. Discussions about continuation, a smaller scope or additional capital must take place before that point.
An operating reserve does not guarantee survival. Its purpose is to allow time to select and implement an alternative. Bank balances also differ from genuinely usable resources: money committed to existing obligations does not provide unrestricted economic flexibility. The calendar should connect each product milestone and commercial finding with the cash expected to remain when management can act on the evidence.
Sensitivity analysis identifies the variables that change the conclusion
With monthly outflows held at R$ 120,000, reducing receipts from R$ 40,000 to R$ 20,000 increases net burn to R$ 100,000. The initial R$ 600,000 then lasts six months. If receipts instead rise to R$ 60,000, net burn falls to R$ 60,000 and runway reaches ten months. These calculations assign no probabilities. They show the exposure of cash to one variable.
Spending can also be tested. With receipts of R$ 40,000, an actual R$ 20,000 reduction in monthly outflows lowers net burn to R$ 60,000, producing the same ten-month runway. Arithmetic equivalence does not imply operational equivalence. Cutting an activity that supports sales may reduce later receipts; postponing a hire may jeopardize work required for contract renewals. The saving must be assessed alongside its consequences.
Sensitivity analysis must respect constraints. It makes no sense to assume immediate savings on the entire program contract if installments remain obligatory or switching creates costs. Equally, increasing sales without allowing for implementation and support can manufacture an improvement that does not exist. A variable should be changed independently only while that independence remains defensible. Beyond that point, a combined scenario is necessary.
For investors, the most useful output is an order of investigation. If small payment delays eliminate decision time, commercial contract quality deserves priority. If runway depends on replacing services without additional cost, transferability becomes central. Analysis should direct diligence and negotiation. It cannot turn a spreadsheet that reacts strongly to uncertain inputs into a reliable forecast merely by adding more decimal places.
Combined scenarios should describe mechanisms of loss
A coherent downside scenario begins with an identifiable cause. Suppose a product integration is delayed and some customers postpone payment, reducing monthly receipts to R$ 20,000. Corrective work simultaneously costs an extra R$ 20,000 each month, increasing outflows from R$ 120,000 to R$ 140,000. Net burn becomes R$ 120,000, so the initial R$ 600,000 lasts five months. Lower receipts and higher expenditure arise from the same operational problem.
Now add a one-off R$ 30,000 expense in month three for technical handover and recovery of documentation. With no other changes, month one ends at R$ 480,000 and month two at R$ 360,000. Month three uses R$ 150,000 net, leaving R$ 210,000. Month four ends at R$ 90,000, which is consumed three-quarters of the way through month five. Total runway is 4.75 months under the same uniform-flow simplification.
Describing this only as a percentage reduction in revenue would hide its mechanism: delay, postponed receipts, rework and technical handover. Understanding that chain enables discussion of specific controls, including incremental delivery, code access, confirmation of customer dependencies and an alternative provider's availability. Each measure should address an identifiable link, with an owner and a cost. Naming a control is not evidence that it can be executed.
Also test a situation in which the product works but the market does not respond. More development may consume cash without removing the commercial constraint. A suitable response could be narrowing scope, reconsidering the target customer or stopping a workstream. Different scenarios require different actions. A plan that recommends accelerating investment in every situation is not using the information the analysis produces.
Value services without confusing price, cost and benefit
The example's R$ 180,000 service charge needs examination from three perspectives. Price is the contractual amount. Replacement cost is what the company would pay elsewhere for comparable outputs, including coordination and transition. Economic benefit depends on how those outputs improve the business. None automatically substitutes for another, and all may be uncertain while the product is still being built. An expensive service can deliver little transferable benefit.
A useful comparison breaks the package into business preparation and governance, feature development, commercial support and ongoing guidance. For each component, identify the team, expected effort, verifiable output and dependencies. Comparing the entire monthly charge with one professional's pay is insufficient. Accepting a high price merely because the package includes informal access to a network, without defined commitments, is equally weak analysis.
Under the downside scenario, ask how much work would remain useful if the program ended in month three. A documented assessment may retain value; a feature without tests, access or adequate licensing may require reconstruction. Recoverable value does not necessarily track the percentage of hours used. Partial work should be assessed through its transferable usefulness and what another team would need to finish it.
If services are paid for with equity or linked to a convertible instrument, the analysis changes form but remains necessary. Define how compensation is earned, what happens to unaccepted work and the consequences of termination. Ownership arrangements do not assure service quality. Paying for work with economic rights may align interests while also making replacement more complicated, particularly if those rights survive an operational relationship that has ended.
Deliverables and intellectual property determine exit costs
A description such as complete product development leaves too much room for interpretation. The contract should identify functionality, acceptance criteria, documentation, appropriate testing and integration responsibilities. Investors need to understand whether the program supplies an operational system, a prototype, a team available for a specified period or a combination. Each model allocates delivery risk differently, and a demonstration does not establish that all contracted outputs are complete.
The intellectual property chain deserves its own review. Distinguish the provider's pre-existing materials, third-party components and work created for the company. Check how employee and subcontractor rights reach the party proposing to transfer or license them. Payment of a Brazilian service invoice, by itself, does not clarify the entire set of permissions and restrictions. The rights required to operate, modify and maintain the product must be understood.
Continuity also requires practical control. Repositories, operating environments, infrastructure accounts, deployment documentation and administrative access need designated owners and transfer rules. A commercial disagreement should not prevent the company from operating or repairing its product. Security remains relevant: continuity does not mean indiscriminate credential sharing. Access should be individual, limited and revocable, with responsibilities clear enough for a replacement team to work safely.
Simulate termination before signing. What material will be delivered, in which formats, within what period and with what support? Do any licenses expire? Can another provider run the system? These questions make intellectual property a financial variable. An asset lacking sufficient rights or maintainability may require the extraordinary expense modeled in the downside scenario and still prolong the delay. Formal ownership without usable access may provide little immediate operating relief.
Conflicts of interest emerge in the allocation of money
When an investor is connected to the provider, part of its investment may return to the group as service revenue. In the example, R$ 180,000 of the R$ 600,000 would be paid to the program over six months: 30% of the investment. This describes a cash flow, not a conclusion about abuse or quality. It makes the selection process, defensibility of the price and identity of the approvers material questions.
The conflict intensifies if further funding depends on retaining the related provider. A company may accept unsuitable services to avoid losing financing. Incentives may also favor recommending additional development, maintaining a weak project or presenting unfinished work as complete. Diligence should examine compensation, ownership, bonuses and targets for those recommending and approving decisions. An incentive can influence conduct even when nobody has explicitly instructed improper behavior.
Proportionate controls include relationship disclosure, documented comparison of alternatives, abstention by interested parties and approval by a body without a direct benefit from the contract. The design depends on the ownership structure and applicable instruments. A formally independent adviser who depends economically on the program may not provide the expected challenge. Examine the actual relationship rather than relying on a title.
Transparency should extend to how opportunities are allocated among supported companies. Technical teams, commercial introductions and partners' attention are limited resources. If several projects depend on the same people, the program should explain how competing priorities are resolved. For an investor in one company, membership of a larger portfolio does not assure stable access to the resources assumed in that company's projections.
Useful evidence measures progress without anticipating conclusions
Effective monitoring combines technical, commercial and financial evidence. A provider's prepared demonstration may be an initial technical signal, but should be supplemented by acceptance criteria and use appropriate to the product's stage. Meetings and proposals show commercial activity; contracts, implementation and receipts demonstrate different advances. Financial reporting should reconcile actual cash with supporting documents and future commitments rather than report a balance without its obligations.
Indicators should address the uncertainty that matters. If the thesis depends on reducing implementation costs, counting delivered features is insufficient: examine effort per customer, rework and specialist intervention. If it depends on repeatable sales, a customer won through a founder's personal relationship can be valuable without proving a reproducible sales process. The evidence may support a narrower conclusion than the headline metric suggests.
Every report should separate observation from inference. A useful formulation is: two customers have paid; three others are negotiating; the model assumes one will start paying next period. Investors can then test the assumption without disputing the facts. Counting all five as established revenue instead converts conversion risk into a later surprise. Clear classification also makes it easier to explain why actual receipts differ from an earlier projection.
Evidence collection has costs and privacy limits. Access to customer documents needs an appropriate purpose, permission and protection. Full personal records need not circulate merely to substantiate a metric. Samples, aggregated information and controlled review may suffice. The program should produce documentation capable of supporting decisions without exposing the company or its customers unnecessarily. More extensive access is not automatically more useful evidence.
Conditional installments change financing risk
Consider a variation: the intended investment remains R$ 600,000, but only R$ 300,000 arrives initially. The other R$ 300,000 would be released at the beginning of month four after a contractual milestone. At R$ 80,000 of monthly net burn, the company ends its first three months with R$ 220,000, R$ 140,000 and R$ 60,000 respectively. It approaches milestone assessment with little room for delay or disagreement.
If the second installment arrives on schedule at the start of month four, cash rises from R$ 60,000 to R$ 360,000. After that month's burn, R$ 280,000 remains. The path then matches the fully funded base case, with exhaustion halfway through month eight. If the installment does not arrive, R$ 60,000 covers only three-quarters of a month. Total runway falls to 3.75 months.
Investors may regard staged disbursement as protection, but it creates interruption risk for the asset they are funding. A subjective milestone, or one dependent on a third party, may block release despite substantial work. Define completion evidence, the verifier, response deadlines, dispute treatment and any opportunity to remedy shortcomings. The company should not plan a discretionary installment as though it were already available cash.
Avoid circular dependencies: requiring revenue that depends on a deliverable whose cost depends on the withheld installment. A milestone must be achievable using resources already committed. Instruments and ownership arrangements must reflect actual disbursements and agreed conditions; this arithmetic example does not settle legal documentation. Protection for one party becomes less effective if it makes the other party's expected performance unachievable.
Negotiate disagreements before cash dictates the answer
Parties may agree on the total expenditure while disagreeing about its usefulness. The program might favor two more months of development, management might prefer selling a narrower solution, and the investor might want to preserve cash. Productive negotiation identifies the assumption each option would test and the money consumed before enough evidence appears to support a decision. The relevant comparison includes the cost of learning.
In the base case, suppose disagreement arises at the end of month four, with R$ 280,000 remaining. A two-month test maintaining R$ 80,000 monthly net burn uses R$ 160,000 and leaves R$ 120,000. That falls below the hypothetical R$ 160,000 reserve. Approving the test without changing the budget, duration or financing therefore means knowingly accepting less room to react. The reserve should influence the decision before the test begins.
One alternative is to narrow the experimental question: test an integration with one customer rather than expand the product across several segments. Savings must be demonstrable, with unavoidable obligations retained in the spreadsheet. Renegotiating scope and payment is another possibility, but a change becomes an established fact only after acceptance. Sending a proposal to the provider does not change the cash contractually committed.
The decision record should identify options considered, available evidence, people excluded because of conflicts and the criterion for ending the test. It should also state who can change direction after a negative result. This reduces later disputes about authorization. Disagreement about the future is normal; concealing it inside a single projection prevents investors from understanding the risk they are accepting.
Assign operating responsibilities and correct modeling errors
Company management should maintain the budget and remain accountable for operations. The provider reports deliverables, effort, obstacles and transition obligations. Investors exercise information rights and participate in decisions established by their instruments. A designated financial function reconciles cash and commitments, while a sufficiently independent technical review helps assess acceptance and continuity. These roles should fit the company's scale without leaving critical responsibilities unassigned.
A frequent mistake is adding advertised services to available cash. Correct it by rebuilding flows by type: money received, expenses paid, conditional discounts and work still to be delivered. Another mistake is treating a commercial contract as an immediate receipt. Correct that with a calendar covering implementation, acceptance, invoicing and payment terms, including possible disputes or cancellation. The signature and the cash arrival represent different events.
Expenses can also be counted twice: the program's R$ 180,000 is included in monthly spending and then deducted again as an upfront cost. That would understate available cash in this example. The opposite error is excluding the contract as investor support even though the company must pay it. Reconciliation should allow every amount to be traced from its contractual clause through the projection to the actual payment.
When a material error is discovered, silently replacing the spreadsheet is insufficient. Explain the correction, recalculate affected scenarios and revisit decisions based on the earlier version. The purpose is not to punish every disappointed estimate, but to distinguish a revised assumption, a calculation failure and omitted information. That distinction guides whether to improve the model, strengthen controls or reconsider confidence in the counterparty.
Legal structure limits what a projection can promise
Financial analysis must correspond to the rights acquired. Equity does not itself create an obligation to repay the investment. Loans, convertibles and receivables transactions have distinct terms requiring review of their instruments and applicable structures. Existing eligible receivables differ from sales the company hopes to make. A commercial projection may support a business hypothesis, but cannot turn an expectation into an established receivable.
The Brazilian organizational form matters. A MEI, or individual microentrepreneur, cannot admit equity partners while remaining a MEI; any proposed ownership investment must address that incompatibility. A donation to a nonprofit entity is not an equity investment and must not promise a financial return. Broad language about support or acceleration should not obscure rights that the chosen structure does not provide.
Public securities distribution and regulated financial activities require appropriate structures and providers. A technology interface, investment language in a contract or a CNAE economic activity classification does not automatically confer permission. Public authorities' materials can provide reference entry points for study; they do not establish a program's or platform's authorization for a specific transaction. Legal review must examine the actual activity and parties in the Brazilian context.
KEZS connects entrepreneurial projects and potential funders through technology and intermediation. Investors conduct their own diligence and make their own decisions; KEZS does not assign a credit score. Transparent documentation may help compare assumptions and rights without guaranteeing approval, repayment, profitability, liquidity or capital preservation. Blockchain integrity does not establish information's truth, legal title, reserve solvency or payment certainty. Private documents should not be placed on public blockchains.
Finish with a decision that can be implemented
Diligence should answer four questions: which assumption justifies the investment, what loss may follow if it fails, how much time is available to respond and who can execute that response? The answer must connect product, cash and contracts. A program's experience is insufficient if the company relies on uncommitted resources or deliverables it cannot transfer. A response plan also needs feasible costs and responsibilities.
In the hypothetical case, full funding supports 7.5 months in the base case; lower receipts reduce that to six months; delay with additional monthly costs produces five months; and the extraordinary expense reduces the last scenario to 4.75 months. In the installment variation, failure to release the second part leaves only 3.75 months. These are distinct conditional outcomes, not simultaneous forecasts or probabilities of survival.
A favorable decision may require specific changes: clearer acceptance criteria, organized technical access, revised service pricing, independent handling of related parties or milestones compatible with available cash. Negotiation also generates information. Resistance to documenting an essential deliverable or disclosing an economic connection can change the assessment as much as a revenue revision. Any proposed control must be assessed for its cost, feasibility and remaining limitations.
The decision may instead be to await specific evidence, reduce the transaction's scope or decline to invest. After investment, monitoring should compare actual and projected performance, explain deviations and update scenarios after material changes. Sensitivity analysis keeps the gap between assumption and fact visible. It can improve decisions under uncertainty while leaving the possibility of partial or total loss and the absence of an available exit.
Frequently asked questions.
Should an accelerator that receives equity provide services without charging separately?
There is no single economic answer. Equity may compensate for risk, access to capabilities or other obligations, while a separate contract pays for execution. Check for overlap: the same deliverable should not be presented as full consideration for equity and charged again without explanation. Assess price, scope, how rights are earned and available alternatives. Separate charges may be defensible, but require transparency about where investment money goes and who approved the purchase. The total arrangement matters more than whether one component is described as free.
How can a specialized service be assessed when comparable quotations are unavailable?
A lack of comparables increases uncertainty without making analysis impossible. Break the work into stages, verify team capability, estimate effort and commission a limited initial deliverable before expanding the commitment. Examine the cost of abandoning the service and reusing completed work. Record the weakness of the pricing reference rather than treating reputation as sufficient proof of economic value. Progressive acceptance and limited exposure help generate evidence. The initial stage should answer a defined question and leave usable material if the relationship ends.
Must every downside scenario have an assigned probability?
No. With little historical evidence, numerical probabilities can imply misleading precision. A scenario can be useful because it exposes vulnerability, even if its frequency cannot be estimated. Explain why the combination is plausible and which decisions it changes. If probabilities are used, explain their basis and avoid hiding extreme losses inside a reassuring average. The ability to withstand a material event deserves separate assessment even when its likelihood is uncertain. A plausible scenario is a decision tool, not a claim that the event will occur.
How can delivery indicators avoid encouraging inappropriate behavior?
Consider what people might do solely to meet the metric. Paying for feature counts may encourage task fragmentation; rewarding meetings may generate unsuitable contacts; using contracted revenue alone may encourage difficult-to-perform commitments. Combine outcomes, quality and financial consequences. Technical acceptance might consider functionality, documentation and maintainability together. Establish criteria before work starts and revise them if they distort behavior, preserving the history of changes. A metric should help assess the underlying uncertainty rather than reward activity that leaves that uncertainty unresolved.
Does a guarantee or buyback clause resolve the risk of loss?
Not by itself. Examine who assumes the obligation, its conditions, the resources supporting it and the process for enforcement. A buyback promise from a party financially dependent on the same business may fail precisely when needed most. Guarantees also involve formalities, priority, recovery costs and time. Depending on a valid structure, they can modify risk, but should not be counted as available cash or certain full recovery. The promise must be evaluated under the same adverse conditions that would make an investor seek enforcement.
When should technical review be independent of the program?
Independence becomes especially important when a technical conclusion determines capital release, substantial compensation or acceptance of work delivered by a related party. The reviewer needs appropriate expertise and sufficient access, with information protected. The role is to reduce dependence on the assessment of the party receiving payment, not to promise that no defects exist. For a small company, a focused review of critical issues may be more useful and affordable than an extensive audit unrelated to the pending decision.
How can investors distinguish a repairable problem from a failed investment thesis?
A problem is more repairable when its cause is identifiable, the intervention fits available cash and evidence shows customers still want the core outcome. The thesis is more compromised when repeated corrections leave behavior unchanged or revenue depends on conditions already shown to be unworkable. Compare the cost of another test with the additional information expected. Stopping a workstream may preserve resources, but also carries obligations, losses and effects on staff and customers. Neither persistence nor termination is automatically the better response.
How can investors consider others' participation without mistaking it for validation?
Other investors' presence shows interest, but does not automatically establish diligence quality or equal terms. Their rights, time horizons, commercial relationships and objectives may differ. Identify shared assumptions, which documents each investor reviewed and dependencies between decisions. If everyone is waiting for everyone else's confirmation, apparent validation may be circular. Each investor remains responsible for understanding their instrument, conflicts and relevant loss scenarios. Collective participation does not create a repayment obligation or make an eventual exit available.
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.
