Corporate venture capital brings together two objectives that do not always move in the same direction: building the value of an equity stake and solving problems for a sponsoring corporation. The relationship can provide access to specialists, facilities and customers, but it can also create pressure to describe any purchase as a strategic investment. The first governance discipline is to separate three decisions: testing a solution, buying a product and taking capital risk. Each needs its own budget, accountable owner, evidence and agreement.
Corporate investors need to separate three decisions
A pilot can show that a technology works in a particular process without proving that the startup can sell it profitably to other customers. A purchase can save the corporation money without increasing the value of its equity stake. An investment may be defensible even when the business unit is not yet ready to become a customer. Combining these decisions lets one department's commercial results conceal portfolio losses. It can also push the startup into customer-specific work that does little to strengthen its underlying product.
The investment thesis should explain how value is expected to be created and which results will be measured separately. Technical learning, lower operating costs, market access and equity appreciation are possible objectives, but none automatically substitutes for another. If the corporation is willing to pay for learning, that purpose belongs in the budget and the program's evaluation criteria. This separation helps investors recognize a useful experiment, a sound commercial relationship and a promising equity asset, including situations in which only one of those conditions exists.
A workable thesis turns interest into verifiable criteria
A thesis such as investing in artificial intelligence for industry is too broad to guide approval authority. An operational thesis identifies the economic problem, the setting in which the solution will be used, the capability the corporation lacks and acceptable integration limits. It might target solutions that reduce equipment downtime at facilities with sufficient historical data, without requiring unrestricted access to core systems. These criteria allow the team to reject attractive opportunities outside its mandate before consuming months of specialist time.
The relationship between strategic fit and financial potential also needs to be explicit. One approach is to assess them in separate reports, each supported by evidence and stated uncertainties. Strategic fit may depend on the importance of the problem and a business unit's ability to run a test. Financial potential may depend on the accessible market, margins, future capital requirements and sales outside the corporate group. A single combined score can hide a weakness that should block the transaction by allowing an excellent result elsewhere to offset it.
Central assumptions need clear labels. Saying that factory access will shorten product development remains a hypothesis until an observable comparison supports it. A technical team's availability, confirmed by its manager with working hours reserved, is a narrower documented fact. For each important hypothesis, the committee should identify the evidence expected, who will collect it and which decision depends on it. This prevents a strategic justification from surviving indefinitely without being tested and makes clear what would cause the team to revise its recommendation.
Internal sponsorship requires execution capacity and continuity
The internal sponsor explains the business unit's problem, provides access to the right people and takes responsibility for the commercial resources committed. A name on a presentation is insufficient. The commitment should identify who will supply data, supervise the test, authorize process changes and fund the work. When a pilot depends on a particular facility, agreement from that facility's manager matters as much as enthusiasm from the innovation department. Access promised without the operational team's support may never become usable access.
The sponsor also needs defined limits. They may recommend a pilot and confirm whether its results are operationally useful, but should not turn that recommendation into investment approval on their own. Their incentive may be to obtain an inexpensive solution or secure the startup's attention for their department. The investment team must assess whether that priority undermines a business that can serve other customers. Governance should retain the sponsor's practical contribution while preventing one operational perspective from controlling every stage.
Continuity needs its own arrangements: a named replacement, accessible documentation and a rule for changes in priorities. If the sponsor leaves, the pilot may continue only for the period needed for review with the new owner, subject to the contract and existing obligations. The equity stake survives the corporation's reorganization. Before investing, the committee should therefore ask whether the startup still has a defensible financial thesis if it loses this sponsor, how much rebuilding the relationship would cost and who would manage that transition.
Approval authority must address powers, amounts and exceptions
A useful approval matrix links each action to the competent authority and the risk it creates. Signing a confidentiality agreement, releasing data, commissioning a pilot, investing capital and waiving a shareholder right are different actions. Financial value is only one dimension: a small pilot may require sensitive access or transfer valuable intellectual property. The matrix should combine spending, the nature of the obligation, data exposure and effects on rights, instead of classifying everything solely by the amount paid.
One possible arrangement gives the business unit authority over pilot spending, the investment team responsibility for the analysis and an investment committee authority over equity decisions. Legal, security and privacy specialists address matters within their remit. Each assessment should distinguish a prohibition, a mandatory condition and a recommendation. Without that distinction, a technical observation can become an informal veto, while an essential requirement can be treated as an optional comment to speed up closing.
Exceptions need an owner, a justification and an expiry condition. If a test must begin with temporary integration, the record should define permitted data, compensating controls and the event that ends the authorization. Artificially splitting contracts to fit a lower approval limit should be prohibited. Spending connected to the same project must be reviewed together, including foreseeable amendments. Otherwise, a substantial obligation can emerge through a series of small approvals without anyone evaluating the complete commitment or its cumulative exposure.
The committee must leave a decision that can be reconstructed
A committee's quality is reflected in whether someone can reconstruct why it decided as it did using the information available at the time. Its materials should cover the thesis, instrument, economic assessment, commercial dependencies, declared conflicts and adverse scenarios. They should also compare alternatives: invest now, test first, buy without investing or walk away. This comparison helps prevent the work already spent on an opportunity from making investment feel inevitable, even when another route would better serve the approved objectives.
Access to internal specialists is valuable when questions are specific and conclusions remain independent. An engineer can examine architecture, an operations leader can assess implementation and procurement can investigate total cost. These opinions do not replace diligence on the company and its rights. Specialists should disclose limitations, their relationship with the sponsor and their reliance on supplied information. If external advisers are paid for analysis, the scope and payment arrangements should be visible so the committee can evaluate possible incentives.
Minutes should record the decision, conditions, material disagreements and who will verify compliance. Approval subject to vague further clarification postpones the real judgment until after the meeting. A verifiable condition might require evidence of the software's chain of ownership before funds are invested, or restrict approval to a stated instrument and amount. If satisfying a condition changes the substance of the transaction, it must return to the appropriate authority. Records should preserve accountability and meaningful dissent rather than manufacture an appearance of unanimity.
Conflicts of interest arise through everyday incentives
Conflicts extend beyond an executive holding an undisclosed stake in the startup. They also arise when a sponsor wants to report savings, an innovation team wants more pilots and a portfolio manager wants to defend an earlier valuation. One transaction can improve all those indicators while draining the startup's cash and reducing its prospects with other customers. Controls begin by identifying these incentives without assuming that every disagreement reflects bad faith. The issue is whether an incentive distorts a decision or the evidence presented.
Anyone influencing the decision should disclose personal relationships, economic interests, parallel contracts and expectations of future employment. The disclosure must describe the connection well enough to support a proportionate response. Depending on the circumstances, that may mean withdrawal from voting and negotiations, independent review or restricted access. Merely disclosing a conflict while leaving the same person in control of every stage does not resolve it. The response should address the person's actual influence over selection, terms, information and approval.
Suppose the sponsor's bonus depends on pilot savings and they pressure the startup to accept extensive exclusivity. The committee should examine both the corporation's savings and the startup's potential loss of market access. Commercial negotiations may continue, but their outcome should not be presented as an independent investment assessment. If an executive connected to the startup is involved, their withdrawal should be recorded. The purpose is to make decisions defensible when interests overlap, including to other shareholders who may bear the cost of concessions.
Pilots need a hypothesis, a baseline and an exit
A well-designed pilot answers a question capable of changing a decision. Testing whether a system reduces inspection time requires a comparable process, a minimum quality standard and a definition of included costs. The baseline should reflect actual operating conditions, including failures and rework. A test using only new equipment or an exceptional team has limited relevance to ordinary operations. Its report should explain where the evidence applies and where uncertainty remains, rather than extending a favorable result to settings that were never tested.
Advancement criteria should be agreed before results arrive. They can combine technical performance, integration costs, security and the human effort required to maintain the solution. Faster processing may have little value if it requires constant specialist supervision. The sponsor assesses operational usefulness, while the investment team asks whether the result strengthens a product that can be sold repeatedly. Local success requiring extensive customization may support a service business without supporting the valuation assumptions used for a repeatable software product.
The exit is part of the pilot design: delivery of findings, return or deletion of data, treatment of materials and payment for contracted work. A paid pilot does not guarantee a subsequent purchase, investment or renewal. Any next agreement requires its own approval. An extension should identify a new question and an additional budget to prevent experiments from continuing indefinitely. Repeating a test without specifying what remains unknown can turn the startup into a supplier subsidized by the expectation of investment, while consuming capacity it could use elsewhere.
Negotiate intellectual property before collaboration begins
Negotiation starts by separating what each party already owns from what will be developed during the relationship. Existing code, models, authorized datasets, methods and documentation need treatment distinct from pilot-specific results. The corporation may require a license sufficient to operate the solution; the startup may need to preserve its ability to sell the product. Assigning every creation related to the project to the customer can eliminate the very asset supporting the equity investment. The commercial agreement and investment analysis must therefore be read together.
A broad ownership clause does not settle the underlying rights. Diligence should examine agreements with founders, employees and contractors, along with third-party components and their licenses. Investing does not cure incompatible or unauthorized material in the product. The parties must also distinguish corporate data, derived results and general knowledge acquired by the team. A vague definition of derived material could give one party control over improvements that disclose no confidential information, making later product development or sales unnecessarily uncertain.
For joint development, agree who can exploit each result, in which markets and under what restrictions. Any exclusivity should be limited and priced with reference to the market the startup gives up. Rights supporting operational continuity if the supplier closes can be negotiated under specific conditions without granting unrestricted ownership of the product. Legal analysis should translate these terms into economic consequences for the startup and other investors. Both sides need to understand what remains usable, transferable and commercially available after the pilot ends.
Commercial independence protects the startup's market
A corporate investor can open doors while making the startup less attractive to the group's competitors. That perception deserves attention even without an explicit sales restriction. A prospective customer may fear that pricing, architecture or implementation plans will reach the corporate shareholder. Governance should separate information received for investment oversight from customers' commercial information, with access limited by purpose and role. Practical safeguards matter because a contractual promise alone may not reassure customers if information circulates widely inside the group.
Exclusivity, purchasing preferences and rights over a company sale have different effects. Segment exclusivity restricts potential revenue; a preference right can delay negotiations; a right to match an offer may deter buyers unwilling to fund diligence that helps someone else acquire the business. Effects depend on wording and context. Investors should assess the economic consideration received and the impact on future rounds instead of accumulating rights simply because they have bargaining power. A right that appears protective can make future financing or an exit harder.
Procurement must retain its own criteria. A CVC stake should neither guarantee purchases nor waive supplier assessment, and it should not require every group request to be served at a discount. A commercial disagreement should likewise not automatically trigger a threat to withhold an investment already contractually committed. Separate contracts and owners help, but actual behavior must be checked. The startup needs a channel for reporting pressure and obtaining review without relying exclusively on the sponsor responsible for the disputed demand.
A numerical example separates investment, revenue and benefit
Consider a fictional startup, Oficina Dados, and an industrial CVC. Every figure in this example is hypothetical arithmetic, not a market price or expected return. The CVC invests R$ 2 million in a primary round at a pre-money valuation of R$ 8 million. The post-money valuation is R$ 10 million, giving the investor 20% ownership under a simple structure with no other instruments, options or adjustments. The funds enter the startup; the existing shareholders do not sell their holdings in this example.
Separately, an industrial business unit commissions a pilot for R$ 240,000. The startup estimates incremental delivery costs of R$ 180,000. The R$ 60,000 difference is a contribution before the company's other expenses, taxes and other effects; it is not net profit. For the corporation, the test indicates a gross operational benefit of R$ 360,000 over the observed period. Subtracting the pilot price and R$ 40,000 of incremental internal costs produces a simplified net benefit of R$ 80,000: R$ 360,000 minus R$ 240,000 minus R$ 40,000.
That R$ 80,000 belongs in the industrial unit's project economics; it is not a profit distribution to the startup's shareholder. The R$ 240,000 is contractual revenue for the startup and creates no automatic obligation to repay the R$ 2 million equity investment. Nor does the observed benefit justify projecting identical gains across every factory. The committee receives three separate records: the ownership acquired, the pilot's operational economics and the supplier's costs. A useful pilot can then be recognized even if the investment later loses value.
The cash schedule exposes hidden dependencies
Continue the example with R$ 300,000 of opening cash at the startup. The R$ 2 million investment arrives at the beginning of month one. Base operating cash outflows are R$ 250,000 a month for six months. The pilot is paid in three installments of R$ 80,000 in months one, three and five, with incremental costs of R$ 60,000 in each of those same months. For simplicity, assume no other inflows, outflows, taxes or financing effects. Closing cash after six months is R$ 860,000.
The path matters as much as the total. At the end of month one, cash is R$ 2.07 million: R$ 300,000 plus R$ 2 million plus R$ 80,000, less R$ 250,000 and R$ 60,000. It falls to R$ 1.82 million in month two, R$ 1.59 million in month three, R$ 1.34 million in month four, R$ 1.11 million in month five and R$ 860,000 in month six. Total available resources are R$ 2.54 million, while total outflows are R$ 1.68 million. Those outflows comprise R$ 1.5 million of base operations and R$ 180,000 of pilot costs.
Now suppose the final pilot installment arrives after month six, with expenses unchanged. Closing cash falls to R$ 780,000, even though a contractual right to payment may still exist. Governance must distinguish revenue, an existing eligible receivable and available cash; eligibility itself requires assessment for the intended transaction. Hoped-for expansion that has not been contracted cannot fill the gap. The CVC should not assume a further investment either: any follow-on reserve needs an appropriate mandate, available resources and a fresh decision under the approval matrix.
Appreciation, dilution and exit require separate calculations
Suppose Oficina Dados subsequently raises R$ 5 million in another primary round at a pre-money valuation of R$ 15 million. The post-money valuation is R$ 20 million. If the CVC does not participate, its 20% stake is multiplied by 15 divided by 20, leaving 15%. At the new round's price, that holding has a reference value of R$ 3 million. This is neither cash received nor a purchase offer or assured exit. The higher reference value and lower ownership percentage can coexist, while liquidity still depends on future events.
If the company is later sold for R$ 40 million available entirely to shareholders, 15% would correspond to R$ 6 million. This calculation assumes no debt deduction, preferences, transaction expenses or other adjustments. A sale for R$ 8 million under the same assumptions would yield R$ 1.2 million, less than the original R$ 2 million investment. These outcomes illustrate how a useful pilot and a subsequent round at a higher price can still end in an investor loss. An interim valuation does not settle the ultimate financial outcome.
Actual distributions depend on negotiated rights, company obligations and transaction structure. Liquidation preferences and other mechanisms may change the allocation, but cannot create resources when the business lacks sufficient value. Investors should model outcomes for all shareholders and examine the incentives of each class. Protection favoring the CVC may deter new investors or weaken founder alignment. Assessing those effects belongs in the initial decision and every renegotiation, alongside the possibility of losing the invested capital and having no practical route to sell the stake.
Different instruments do not resolve the same uncertainty
Equity exposes an investor to the business's residual value and the rights agreed. Debt creates contractual payment obligations, subject to the borrower's ability to perform and any legally valid security. A convertible requires analysis of conversion mechanics, maturity and what happens if the expected financing round never occurs. These instruments create different rights and risks. None should be selected simply to avoid difficult questions about valuation, technological uncertainty or cash availability, and a payment obligation is not proof that payment will occur.
A commercial contract pays for deliverables and allocates supply responsibilities. Advance payment may help the startup's cash position, but exposes the buyer to delivery risk and needs suitable contractual treatment. Existing eligible receivables differ from sales merely hoped for. Donations to nonprofit entities serve a different purpose and are neither equity investments nor promises of financial return. A project organized as a Brazilian MEI cannot admit equity partners while retaining MEI status. The legal form and proposed rights must be compatible before the parties rely on them.
The applicable Brazilian structure must be assessed for the actual transaction. Public securities distribution and regulated financial activities require appropriate arrangements and participants; technology, a standard contract or a CNAE business activity code does not replace authorization. An intermediation platform does not remove investor diligence. Before assessing commercial attractiveness, establish who may perform each act, who is responsible for the offering or service and which operating limits apply. Specialist advice may be necessary, and reference materials alone do not establish permission to conduct a regulated activity.
Useful information requires proportionate access rights
Investors need to monitor cash, performance and material events, but do not need every dataset the startup holds. Periodic reporting might cover cash balances, commitments, customer concentration, product progress and departures from plan. Information rights must take account of commercial confidentiality, personal data and obligations to third parties. A seat on a governance body does not make customer files freely distributable throughout the corporation. The purpose of oversight should determine access, with controls reflecting the sensitivity of the information and the recipient's responsibilities.
Efficient diligence organizes evidence around questions and access levels. Contracts and records support statements about rights; financial statements and authorized bank records support financial analysis; documented tests support technical conclusions within their scope. The team should identify authorship, versions and gaps, distinguishing reported information from corroborated information. A polished file does not establish that its contents are true. Missing essential documents should remain visible as unresolved risk, with clear consequences for the decision rather than an assumption that presentation quality compensates for absent evidence.
KEZS connects entrepreneurial projects and potential funders through technology and intermediation. Organizing analysis by project, intended use of funds and transparent documentation can help investors formulate questions and compare evidence. Investors retain responsibility for diligence and decisions; KEZS does not assign a credit score. Technology records also have limits: blockchain integrity does not establish truth, legal title, reserve solvency or a payment guarantee. Private documents should not be placed on public blockchains. A record's resistance to alteration does not independently verify the underlying transaction or claim.
Adverse scenarios test governance before a crisis
Suppose a pilot meets its technical target, but the corporation suspends the project after a strategic change. The business unit must meet contractual obligations, record the learning and close access in an orderly way. The investment team should reassess the thesis without automatically treating loss of the sponsor as proof that the business is unviable. Key questions include how many independent customers exist, how much of the product depends on that facility and whether available cash allows a move toward another market without emergency financing.
In another scenario, the solution works but a dispute arises over code developed by a contractor. This should trigger legal review of rights and proportionate measures to support continuity and contain exposure. Additional investment should not conceal the unresolved issue. If funding is proposed to settle the dispute, the committee needs to understand its use, cost, alternatives and effects on other shareholders. Preserving an earlier investment does not by itself justify a new one. The decision must assess the new commitment on its own evidence and terms.
Legitimate disagreement is also possible: operations wants to buy, investments recommends against investing and the startup prefers another customer. Keeping only the commercial contract may be the right result. In renegotiation, each concession needs explicit consideration and an impact assessment. Cutting the pilot price for a vague promise of access to the group transfers value without a verifiable commitment. Investors should be able to defend their choice even if a new round, commercial expansion or exit does not occur within the expected horizon.
Correct mistakes and close the governance cycle with discipline
Reports can reveal recurring errors: adding commercial benefits to equity returns, treating future receipts as cash, ignoring internal labor or closing a pilot without a conclusion. Correction begins by reclassifying information while preserving the previous record. If the example's net benefit were reported as R$ 120,000 because R$ 40,000 of internal costs had been omitted, it should be corrected to R$ 80,000 with an explanation of the method. Silently changing the baseline undermines trust and prevents the organization from learning why the error occurred.
Periodic review should bring together the portfolio owner, sponsor and relevant functions to examine deviations, conflicts and upcoming commitments. It need not repeat the entire original approval; it should focus on changes and the decisions they trigger. Late cash receipts, changes in ownership of key assets, customer concentration and departures of key people may require different responses. Each action needs an owner and evidence of completion. This helps prevent an unresolved issue from reappearing in report after report as a comment with no practical consequence.
Before committing capital, investors should be able to explain the financial thesis, strategic usefulness, approval powers, movement of money and rights attached to each instrument. They should also know how the pilot ends, who controls intellectual property and whether the startup can sell outside the group. Where an answer depends on a hypothesis, specify its test and the possible loss if it fails. This discipline makes risk more understandable and decisions easier to hold accountable. It provides no guarantee of approval, repayment, profitability, liquidity or preservation of capital.
Frequently asked questions.
Should the internal sponsor have veto power over the investment?
The sponsor should control their department's resources and commitments within the approval matrix. If they withdraw access to an essential facility, an investment assumption may fail and require review. That differs from a general veto based on personal preference. The equity decision belongs to the authority defined in the mandate, which must understand sponsor dependence and assess alternatives. Separating these responsibilities recognizes an operational constraint without confusing the department's availability with a financial judgment about the startup. Any replacement assumption should be documented rather than silently substituted.
Is an approved and paid pilot enough evidence to invest?
It demonstrates a defined commercial relationship and, with suitable measurement, may establish performance in that setting. Investment still requires analysis of rights, team, cash, market, implementation costs and repeatability across customers. An experimental budget may have funded the pilot, or delivery may depend on services that do not scale. The analysis should state exactly what the test confirmed and which business assumptions remain unresolved. Payment alone does not establish product-market fit, durable margins or an eventual buyer for the investor's stake.
How can exclusivity be negotiated without undermining the startup's value?
Define the product, segment, territory, duration and conditions for maintaining exclusivity, then estimate the opportunities being restricted. Consideration may include verifiable commercial commitments, subject to legal and economic assessment. Broad exclusivity based only on expected access to the corporate group is difficult to price. Review how other customers, future investors and potential acquirers may perceive the restriction. Even a clearly documented agreement can impose substantial costs, so the committee should assess whether its compensation and strategic purpose justify the market access surrendered.
What if the CVC wants information about competitors served by the startup?
Investment oversight rights must respect the startup's obligations and the purpose of the information. Aggregated indicators may meet part of the need without disclosing identifiable customers' contracts, prices or plans. Sensitive matters may require restricted recipients and consideration of independent advisers. If proportionate access cannot resolve the issue, the investor must decide with that limitation rather than pressure the startup into improper disclosure. A shareholder's desire for more detail does not override confidentiality commitments or justify circulating customer information across unrelated corporate functions.
How should an investment be assessed if it saves the corporation money but loses equity value?
Present both outcomes separately, then compare them with the approved mandate. In the example, the pilot's simplified net operational benefit is R$ 80,000; it does not erase a possible loss on the R$ 2 million investment. If the program also funds strategic learning, that objective needs its own budget and criteria. Evaluation can recognize operational usefulness and financial loss together without accounting offsets or narratives that conceal either. The benefit belongs to the business unit's project analysis, while the equity result depends on the investment's rights and eventual outcome.
Who should own improvements developed during the pilot?
The answer depends on each party's contribution, existing rights and intended economic use. The agreement should distinguish pre-existing technology, specific adaptations, data and general improvements, with compatible ownership and licenses. The startup needs to know what it can commercialize afterward; the corporation needs to know what it can continue using. Resolve these questions before development begins, because a later dispute can damage both the commercial operation and the equity stake. Also verify that contributors and third-party licenses support the rights the parties propose to allocate.
When does it make sense to stop the analysis or decline a follow-on round?
When evidence no longer supports the thesis, essential rights remain uncertain or the new balance of risk and price falls outside the mandate. Capital already invested is insufficient reason to invest more. Consider cash, alternatives, dilution and the consequences of not participating, without assuming the corporation will rescue the startup. Documenting a refusal helps distinguish investment discipline from withdrawal caused by internal disorganization. A follow-on reserve is a planning resource subject to its mandate and approvals, rather than an automatic commitment to finance every subsequent need.
What minimum documentation helps review an opportunity's governance?
A useful package includes the thesis and hypotheses, accountable owners and approval powers, conflict disclosures, instrument analysis, ownership structure, technology rights, pilot budget and cash forecast with assumptions. Add acceptance criteria, investment conditions and decision records. The package should be proportionate and does not replace specialist diligence. Its value lies in connecting claims to available evidence and showing who decides, under which conditions and with which risks. Unresolved gaps should be explicit, including who will address them and whether a decision must wait for the missing evidence.
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.
