Governance, Decision Rights and Conflicts of Interest in Accelerators and Venture Builders

An investment assessment begins by separating roles

Accelerator and venture builder describe operating models; they do not fully explain anyone's rights or responsibilities. A mainly educational program needs different controls from an organization that hires developers, manages suppliers and receives compensation linked to equity. Start with the actual activities: who performs each service, who pays, who takes on obligations, who receives economic rights and who can bind the company. A program's commercial name cannot replace this map, and two organizations using the same label may require very different assessments.

This article offers an assessment method using entirely hypothetical examples. All amounts, expenses and ownership percentages illustrate decisions; they are not market prices, observed results or investment recommendations. The purpose is to help investors examine whether work, money and influence translate into verifiable obligations. Useful governance makes clear what each party delivers, what it receives, which decisions it controls and what happens when the plan fails. It also makes the cost of oversight visible, so controls can be proportionate to the project's exposure.

Separate agreements make the economics understandable

Founder preparation, product development, commercial support and investment may form part of one negotiation, but each needs its own terms. A services agreement defines scope, price, acceptance and responsibility. Corporate documents address ownership, dilution and shareholder rights. A convertible instrument sets conditions for possible conversion and any other agreed rights. Debt creates payment obligations. Combining everything under a general promise of partnership prevents investors from identifying what the company has purchased, financed or merely included in its plans.

Providing a service does not automatically amount to contributing invested capital. Mentoring, access to a network and software development can be compensated through negotiated mechanisms, including arrangements connected to equity, but the legal structure must support the arrangement. Recording a consulting fee in a spreadsheet does not make it paid-in capital. Investors need to distinguish a cash payment, an obligation to pay, compensation subject to conditions and an ownership interest that has actually been established. These categories affect the company's resources and obligations differently.

Separation also makes failures easier to address. If the product is not delivered, the parties should be able to dispute the service without making an already completed cash investment uncertain. If the investment does not happen, they need to know whether the services continue, shrink or end. Documents may refer to one another, provided they specify the events connecting their obligations. Hidden dependencies deserve particular attention, such as requiring another purchase before releasing funds previously presented as available. The company must understand the full bargain before committing.

Value services without treating a list price as an investment

Valuation starts with the work the project needs. Investors should be able to identify the professionals involved, their expected commitment, deliverables, assumptions and excluded activities. Two development packages are not comparable simply because both promise a minimum viable product. Integrations, testing, documentation, knowledge transfer and rights to the code can substantially change what is being purchased. Compare proposals on a consistent basis and consider what the company would spend to obtain an equivalent result while retaining the ability to operate independently.

Keep three figures separate: the provider's cost, the agreed price and the economic value the company receives. A provider might incur costs of R$ 108,000, charge R$ 180,000 and create a benefit worth less or more than that price. The difference between cost and price does not establish abuse; it may support staff, overhead and business risk. However, presenting a R$ 240,000 list price as an investment already made, when the agreed scope costs R$ 180,000, overstates the economic contribution and can distort the equity interest being requested.

Discounts also need context. A reduction from a list price rarely used in practice does not demonstrate a benefit for the project. More useful evidence includes comparable proposals, a breakdown of the work, verifiable execution references and explanations for differences in quality or responsibility. Where reliable comparisons are unavailable, staged budgets, independent review and exposure limits can help. Uncertainty should influence price and conditions rather than disappear behind an artificially precise estimate. A negotiated number remains an estimate of value, not proof that equivalent value will be created.

A complete example of services, cash and ownership

Consider a hypothetical company negotiating six months of work with a venture builder. The initial proposal shows a R$ 240,000 list price. After reviewing scope and comparing alternatives, the parties agree on R$ 180,000. The program allocates R$ 30,000 to preparation in the first month, R$ 90,000 to product work in months two through four and R$ 60,000 to commercial support in the final two months. For simplicity, each month contains a R$ 30,000 deliverable, subject to its own acceptance criteria. Equal monthly amounts do not mean the underlying work or risks are identical.

The company will pay R$ 10,000 in cash each month, totaling R$ 60,000. The remaining R$ 20,000 for each stage will be compensation conditional on delivery, creating an accumulated economic reference of up to R$ 120,000 for negotiating equity. This reference is neither deposited cash nor automatic permission to contribute services as capital. The agreement needs a legally appropriate mechanism, clear conditions and a rule for rejected stages. Without acceptance, the corresponding amount should not be treated as an established contribution merely because the provider has submitted an invoice or time has passed.

To illustrate only the final economic split, assume a R$ 3.48 million reference for the founders' position, R$ 120,000 recognized for the builder and a R$ 400,000 cash investment from an investor. The total basis is R$ 4 million: founders hold 87%, the builder 3% and the investor 10%. The arithmetic is 3.48 million, 120,000 and 400,000 divided by 4 million respectively. This assumes all six stages are completed and no other dilutive rights exist. The percentages are a contractual comparison scenario, not a verified valuation or a forecast of investment returns.

The cash schedule reveals risks that ownership percentages hide

Continue the same example. The company starts with R$ 200,000 available. Its other operating expenses, excluding the builder's services, are R$ 70,000 a month. Adding the R$ 10,000 monthly payment to the program produces a monthly cash outflow of R$ 80,000. Assume no revenue or other inflows during the six months and a R$ 400,000 investment arriving at the beginning of month three. The closing balance is R$ 120,000 after month one and R$ 40,000 after month two. These balances describe the assumed payment schedule, not a general runway recommendation.

With the expected investment received, month three closes at R$ 360,000. Subsequent closing balances are R$ 280,000, R$ 200,000 and R$ 120,000. The overall check gives the same result: R$ 200,000 initially available plus R$ 400,000 received, less R$ 480,000 paid. The R$ 120,000 of conditional service compensation does not enter this cash flow because it is not cash. Claiming the company received R$ 520,000 in spendable resources would confuse operating finance with a contractual economic reference. Ownership calculations and liquidity calculations answer different questions.

If the investment is delayed and nothing changes, the projected closing balance for month three becomes negative R$ 40,000. That does not mean the company can operate normally with a negative balance; it means the plan runs out of funding before completing the scheduled payments. Governance should trigger a decision during month two: cut expenditure, suspend a stage or obtain an alternative that is actually contracted. Verbal investor interest belongs among uncertain assumptions, not confirmed inflows. Any revised schedule must also consider when payments fall due within the month, rather than relying only on month-end totals.

Decision authority should reflect exposure, duration and relationships

An approval matrix defines who may decide, under which conditions and with what record. Its design should reflect available cash, the duration of commitments and related-party involvement. In the example, management might be allowed to make budgeted purchases up to R$ 8,000 per engagement, with a second approval required above that amount. Yet a commitment of R$ 4,000 a month for twelve months totals R$ 48,000. Assessing only the monthly installment would bypass the purpose of the control. Approval should address the obligation being assumed, not merely the next payment.

These thresholds are hypothetical choices, not recommended standards for every company. The principle is to examine total exposure, including automatic renewals, termination penalties, exclusivity and technical dependence. Hiring the builder itself, or a business connected to its owners, should follow a specific process even below the ordinary approval threshold. The conflict arises from the economic relationship and influence over the choice; splitting the expenditure into smaller contracts does not remove it. A lower payment can still lock the company into a costly or difficult-to-replace arrangement.

An investor may reserve consent rights for material changes, such as related-party transactions, transfers of essential assets or commitments outside the budget. Giving the investor a veto over every purchase can stall operations and blur responsibility. A useful matrix preserves management's ability to execute, sets response deadlines and distinguishes notification, consultation and approval. It should also address genuine emergencies through limited exposure, a recorded justification and later review by someone without an interest in the outcome. An emergency process should resolve urgency while preserving accountability for the decision.

Identify who delivers, who accepts and who oversees

Operational governance must identify actual people and responsibilities. The builder may appoint a delivery lead while the company names a product owner and a budget owner. Technical acceptance should rest on evidence and may require independent assessment for complex work. It is insufficient for one person to write the scope, declare their own work complete and authorize the associated compensation. The greater the economic rights at stake, the more valuable this separation becomes. Responsibility should remain clear even when a small team has to combine some roles.

A committee might include a founder, an investor representative and an independent member with defined duties. Creating a committee does not remove the responsibilities of company administrators or guarantee sound judgment. Its rules should address access to information, recording dissent, replacing members and handling conflicts. If a builder representative also serves on a governance body, everyone should understand when that person acts as a service provider, when they exercise a governance role and which decisions require them to step aside. A title alone does not resolve competing duties.

A practical routine can combine weekly execution reviews, monthly budget reviews and assessments at specific milestones. Each meeting should produce concise decisions identifying the problem, options considered, owner, deadline and financial effect. Long reports that leave issues unresolved create only the appearance of control. Investors need to know who must respond when a risk emerges. A delayed delivery should not remain an unresolved exchange between supplier and founder while conditional compensation continues to be recognized simply because another month has elapsed.

Deliverables need verifiable acceptance criteria

Preparation, product development and commercial support call for different evidence. Preparation might produce a revised budget, a risk map and a documented commercial thesis with identified assumptions. Product work might require demonstrable functions, agreed tests, an accessible repository and deployment documentation. Commercial support might involve account research, qualified contacts and recorded experiments. Attending meetings shows effort, but does not by itself establish that any of these outcomes has been completed. Evidence should match the promised service and allow another competent person to assess it.

In the example, each R$ 30,000 stage needs an attachment specifying deliverables, dependencies and approval criteria. A feature may depend on client data or decisions, so the agreement should explain how delays in those inputs affect deadlines. Distinguish defect correction, scope changes and desired enhancements. Otherwise, the supplier may charge again for work already owed, while the company may demand additional work without acknowledging its budget effect. Clear criteria also help identify whether a delay originates in the builder's work, the company's missing input or an agreed change.

Acceptance may be full, partial or refused, with reasons and a correction period. If a product stage fails an essential test, its conditional compensation should not become earned solely because the schedule has advanced. The opposite problem also needs protection: a vague objection used to withhold payment for demonstrably completed work. Technical review based on evidence and a defined deadline can reduce the cost of deadlock. Investors gain a clearer view of obligations already fulfilled and exposure that can still be limited, renegotiated or stopped.

Intellectual property forms part of the asset supporting the decision

Investors need to understand what the company will control at the end of the program. Developed code, visual identity, models, documentation and databases can have different origins and legal treatment. Separate the builder's pre-existing materials, project-specific creations and third-party components. Contracts should address the rights needed to use, maintain and develop each element. Paying for development does not automatically settle every ownership or licensing question. A working demonstration is therefore only part of the evidence: the company also needs the rights required for its intended operation.

When the builder reuses its own software library, the company may receive a license rather than exclusivity. This can reduce cost and delivery time if the license is sufficient for the business and disclosed to the investor. Review supplier dependence, continuity after termination, transferability and future charges. Maintain an inventory of third-party licenses and their obligations. A functioning product can contain restrictions incompatible with commercialization or a later transaction. The assessment should connect these restrictions to the company's actual business plans instead of treating every license as interchangeable.

The chain of rights also runs through employees and subcontractors. General assurances from the builder do not replace evidence that it can grant the rights promised. Repositories, domains and essential accounts need organized administration, recoverable company access and clear responsibilities. Personal data requires appropriate purposes, access controls and protection; it does not become unrestricted property of whoever built the system. On exit, the handover should include documentation, operational knowledge and credentials transferred through a secure procedure, allowing continuity without exposing private information.

Address conflicts before the vote

A conflict exists when someone's own interest may influence a decision made on behalf of another party. It can arise in supplier selection, service assessment, investor introductions or the allocation of opportunities among client companies. A builder might receive a partner commission, own an interest in a competitor or favor a solution that reuses its infrastructure. None of these relationships alone establishes wrongdoing. The concern is making decisions without disclosing the connection and assessing its effect. Disclosure should be specific enough for others to understand the possible incentive.

The process begins with a clear statement of the relationship, who benefits and how the benefit might affect the choice. It then requires comparable alternatives, assessment by people without a direct interest and a recorded decision. Where necessary, recusal should cover preparation and influence over the process, not just the final vote. If the interested person alone chooses the options presented, leaving the room for the vote may achieve little. Oversight needs to examine how the decision was shaped as well as who formally approved it.

Suppose the builder recommends a tool supplied by a company in which its owners hold an interest. Compare total cost, suitability, data access and the ability to replace it. If selected, record the reasons and applicable controls. One startup's confidential information must not inform another's strategy without an appropriate basis and relevant authorization. Access barriers, separate teams and access logs can help, but some conflicts cannot be managed reliably and require declining the work. The cost of separation and the consequences of a failure should be part of the assessment.

Commercial incentives can shift the program's objective

Service fees, equity and fundraising commissions create different incentives. A provider paid by the hour may have an incentive to extend scope; an equity recipient may prioritize a rising valuation narrative; someone paid when a transaction closes may press for a premature deal. These risks do not determine how any organization behaves, but they help investors choose controls. Identify who benefits from each decision and who bears the cost if it proves unsuitable. A compensation structure should be understood through its possible effects on behavior, not only its headline price.

Commercial support metrics should connect activity with verifiable outcomes. A long contact list is not the same as qualified opportunities; an opportunity is not a signed contract; a contract is not collected cash. Variable compensation should define how cancellations, defaults, refunds and sales attribution are handled. Where several channels contribute, the rule should prevent the builder from earning commission on all company revenue merely because it offered initial advice. The agreement needs a defensible link between the compensated work and the revenue used to calculate the fee.

The connection between services and fundraising also needs boundaries. The company should know whether it can reject an introduced investor, hire another adviser or end the support without disproportionate penalties. Exclusivity may have an operational rationale, but needs a defined scope, duration and corresponding commitment. Investors should question metrics rewarding only closing speed or announced volume. A transaction can look successful for an intermediary while leaving the company with obligations its cash and governance cannot support. Closing a deal does not establish that its terms serve the business.

Evidence makes disagreement possible without relying on authority

Due diligence should reconstruct the connection between proposal, execution, acceptance and compensation. Organize contracts, budgets, delivery evidence, decisions and the capitalization table in identifiable versions. A project-oriented platform can potentially help present the problem, use of funds, milestones and documents coherently. This organization can reduce search costs, but cannot turn statements into facts. Investors remain responsible for investigating inconsistencies, seeking specialist advice where appropriate and reaching their own decisions. A polished presentation should make the underlying evidence easier to examine, not replace it.

Classify information as a documented fact, a party's statement or a planning assumption. Evidence of a completed bank transfer differs from an intention to invest. A test in a controlled environment differs from a capability demonstrated in operation. A sales estimate prepared by the commercial team remains an assumption until appropriate evidence of contracting and collection exists. This classification helps prevent projections from gaining the appearance of historical performance through repetition. It also makes updates more useful: a new piece of evidence may change an assumption's status without validating every other claim in the plan.

KEZS connects entrepreneurial projects and potential funders through technology and intermediation. It does not assign a credit score or replace investor due diligence. Its potential usefulness lies in making documents and assumptions easier to understand; any claim about an operational capability requires specific evidence. Even an intact record can contain incorrect information. Review should therefore examine origin, consistency and context while controlling access to private documents. Neither the presence of a project on a platform nor the organization of its materials guarantees approval, repayment, profitability, liquidity or preservation of capital.

Resolve disagreements without dissolving responsibility

Founders, builders and investors can disagree in good faith about price, priorities and risk. A founder may consider an integration essential, the builder may see additional scope and the investor may prefer preserving cash. Discussion improves when each party states its assumption, evidence and the cost of being wrong. Identify the decision's purpose: fulfilling an existing obligation, testing demand or building a future capability. Without that framing, personal preferences may be presented as unavoidable technical requirements, making it harder to compare real alternatives and assign responsibility.

In the example, suppose a change in month three requires another R$ 24,000 in cash. Approving it without an offset would reduce final cash from R$ 120,000 to R$ 96,000, with all other assumptions unchanged. Options include removing another item, deferring the change or funding the additional work through a contracted source. The R$ 120,000 reference for conditional compensation should not increase silently. Any price or ownership revision needs its own record and updates to dependent documents. The decision should show both what the change adds and what resources it consumes.

A deadlock process can begin with negotiation between the responsible people, move to technical review and reach the body authorized to make the economic decision. Deadlines prevent the dispute from consuming the project's continuity. Interim measures should preserve cash, code and data without prejudging who is right. Investors do not need to arbitrate engineering details to require a workable decision process. Recorded dissent is useful when it identifies genuine options and accountability; unanimity produced by dependence is not evidence of sound governance.

Adverse scenarios should change obligations predictably

If the example program ends after four accepted stages, R$ 120,000 in services will have been recognized: R$ 40,000 paid in cash and R$ 80,000 in conditional compensation. The R$ 40,000 of conditional compensation scheduled for the remaining two months has not been earned. What happens to the recognized R$ 80,000 depends on the agreed contractual mechanism; it should not be presumed to be definitive equity or automatically become debt. Documents need to address termination, rights already earned and amounts still subject to conditions. These distinctions matter before anyone relies on a final ownership percentage.

Another scenario is technically correct delivery without customer adoption. If the agreed service was development and experimentation, a lack of sales does not by itself establish breach. It also does not justify calling the program an economic success. Separate a fulfilled service obligation, a rejected commercial hypothesis and a financial outcome. If the builder promised a specific result, the assessment changes according to the contract and evidence. Investors should avoid assigning all business risk to the provider while also refusing to accept any amount of effort as sufficient delivery.

Other risks include departure of a key professional, interruption of the builder's operations and disputes among owners. A continuity plan should provide access to completed work, current documentation, replacement arrangements and cooperation during transition. A termination clause has little practical value if nobody can maintain the product afterward. Losses remain possible despite appropriate controls: equity may lose value, debt may go unpaid and assets may be difficult to sell. Governance improves the ability to respond; it does not guarantee capital recovery or a ready buyer for an investor's rights.

Correct errors and respect the limits of each instrument

Correct errors at their source and in every affected record. If a presentation treated the R$ 240,000 list price as a contribution when the agreed price was R$ 180,000, revise the explanation and associated calculations. If a cash flow counted service compensation as an inflow, rebuild it. Identify the corrected version, responsible person, impact and recipients who need the update. Silently editing a document leaves earlier readers relying on an inconsistent understanding. The correction should make clear whether it changes the proposed economics, a decision already made or only the way information was described.

The structure must also respect the selected instrument. Loans, receivables, equity, convertibles and donations create different rights and risks. Hoped-for sales are not existing eligible receivables. A donation to a nonprofit does not grant equity or a promised financial return. A Brazilian individual microentrepreneur, or MEI, cannot admit equity partners while remaining a MEI. Resolve these differences before presenting the economic offer, using appropriate analysis of the actual situation. Renaming an arrangement does not change its substance, and a useful service model does not make every funding instrument suitable for every business.

Public securities distribution and regulated financial activities require appropriate structures and providers; technology or a Brazilian CNAE business activity classification does not grant permission. Materials from Banco Central, CVM, Portal do Empreendedor and ANPD are reference entry points, not evidence that KEZS or another organization is authorized. Blockchain can help establish record integrity, but does not prove truth, legal title, reserve solvency or payment. Private documents should not be placed on public blockchains. Any proposed recordkeeping benefit must be weighed against access control, confidentiality, cost and the limits of what the record can establish.

A practical decision needs entry conditions and follow-up

Before committing resources, investors should be able to explain the arrangement in plain language: how much cash comes in, what is paid for services, which rights each party receives and which conditions remain outstanding. In the example, that means separating the R$ 400,000 investment, R$ 60,000 in payments and R$ 120,000 conditional reference. It also means recognizing that the 87%, 3% and 10% ownership split depends on the stated assumptions. If other rights exist in the actual capitalization table, update the calculation rather than treating the simplified example as a definitive ownership statement.

Entry conditions may include consistent agreements, a demonstrated intellectual property rights chain, a fundable budget and approval of related-party arrangements. Not every outstanding issue blocks a transaction; some can be addressed through deadlines, spending limits or suspension of a stage. What matters is recording the accepted risk and the person responsible for monitoring it. A condition nobody checks before releasing funds offers little control, however prominently it appears in the documents. The evidence required to satisfy it should be clear enough to support an actual release decision.

After the decision, compare actual cash, accepted deliveries, commitments and disclosed conflicts with the plan. Material deviations require action: correct, renegotiate, reduce exposure or stop. Governance costs should be proportionate without abandoning controls over risks that could threaten continuity. Useful documentation lets investors understand the agreement, test its assumptions and act when reality changes. This is an assessment method, not a prescription for personal asset allocation. Contractual and operational clarity cannot remove uncertainty, but can make a decision better informed, accountable and open to revision.

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.