Summary

  • Resolution 201110.130 named Adiel Ayodele Akplogan, Patrisse Deesse, Krishna Seeburn and Christian Fanchette as AFRINIC bank-account signatories, attaching offices to the four names and referring to a delegation-of-authority table.
  • The surviving public page does not display that table. It therefore does not establish whether any person could act alone, which combinations applied, what account classes or transaction types were covered, or when a bank activated or revoked the mandate.
  • Adjacent October 2011 resolutions add bounded context: the same four were to act under an authorisation policy; external transactions above USD 100,000 or equivalent required Board approval; bankers were to receive relevant policies and resolutions; and a one-year Finance and Audit Committee was established.
  • Those are recognisable control components, not proof of implementation or effectiveness. No sealed evidence establishes a transaction, loss, misuse, conflict, bank rejection or benefit involving any of the four people.
  • The proportionate accountability answer is a dated, privacy-safe mandate-control register joined to protected bank and transaction evidence. It should expose roles, effective intervals, control tiers, combination architecture, review dates and supersession without exposing account identifiers, specimen signatures, credentials, balances or raw transactions.

The concrete act

In October 2011, AFRINIC’s Board put four names beside one of the company’s most consequential internal credentials: authority to sign on its bank accounts. The resolution identified Adiel Ayodele Akplogan as CEO; Patrisse Deesse as Director Finance & Accounting; Krishna Seeburn as representative of the Board of Directors in Mauritius; and Christian Fanchette as Director, HR & Administration. It then described those four as bank-account signatories under a delegation of authority “in the table above.”

That final reference is where a simple personnel announcement becomes a governance problem worth examining. The surviving public page names the credential-holders, but it does not display the table that would tell a reader what the credentials could do. Four names can suggest distribution. Four offices can suggest that financial execution was not placed visibly in one pair of hands.

Neither fact answers whether one signature was enough, whether signatures had to be paired, whether a finance officer had to join an executive or Board representative, whether different amounts triggered different combinations, or whether different banking services followed different rules.

The stakes are practical rather than theatrical. A payment mandate sits at the junction between a company’s decision and a bank’s execution. If its rules are precise, current and independently testable, it helps prevent an approved obligation from being paid by an unauthorised route, an obsolete office-holder from retaining access, or a high-value instruction from bypassing the intended approval tier. If the public description is incomplete, outsiders cannot tell whether those protections were designed, even if a complete private mandate may have existed. The correct response is not to fill the silence with accusation.

It is to identify exactly which join is missing and which evidence would close it.

Resolution 201110.130 is unusually useful because it supplies a testable object. “Sound governance” is an aspiration. A signatory slate is a dated control object that can be compared with an authorisation policy, a bank mandate, effective dates, access records and individual transactions. The October record lets readers begin that comparison. It does not let them finish it.

What the resolution fixed—and what it did not

The public act fixed at least three facts. First, AFRINIC published a four-person signatory slate. Second, it attached a contemporaneous office to each person, allowing the slate to be understood as a combination of institutional roles rather than a bare list of private individuals. Third, it expressly linked their signatory status to a delegation-of-authority table.

The offices matter. On the face of the published titles, the slate included the chief executive, the director responsible for finance and accounting, the director responsible for human resources and administration, and a Mauritius-based Board representative. This looks like an attempt to distribute involvement across operational leadership, finance, administration and a Board-facing position. That is the strongest reasonable institutional defence of the design visible on the page. It shows more thought than a generic statement that “authorised officers” could deal with banks.

But role variety is not the same thing as segregation of duties. A title does not prove independence. It does not show who prepared a payment, who verified the payee, who established the underlying obligation, who checked the budget, who supplied supporting documents, who entered the bank instruction, who approved it inside the banking platform, or who reconciled the settlement later. Nor does the presence of a Board representative establish that a Board-level signature was required for every instruction—or for any particular class of instruction.

The role description is evidence of how AFRINIC presented the slate, not evidence of every responsibility carried by each office.

The resolution also did not prove that a bank had received the mandate. A corporate board can authorise named people internally; a bank still has to receive the documentation it requires, validate the signatories, configure access, apply the correct combination and limit rules, and set an effective date. Those steps can occur promptly, later, imperfectly or not at all. The adjacent direction to furnish relevant policies and resolutions to AFRINIC’s bankers shows that the Board recognised a communication step. A direction to communicate is not a receipt, an acknowledgement or an activation record.

Most importantly, signatory status did not itself authorise the underlying economic decision. If a company owes a supplier, pays staff, refunds a customer, meets a tax duty or settles another valid corporate obligation, the authority for that obligation arises upstream. It may come from a contract, an approved procurement, a payroll process, a statutory duty or another properly authorised act. The bank signatory executes an instruction within a control chain. A signature cannot manufacture a missing contract, budget approval or corporate decision merely because the bank accepts it.

This distinction protects every side of the analysis. It prevents AFRINIC from using a signatory resolution as a complete answer to questions about a payment. It also prevents a critic from treating a named signatory as the author of every obligation that may have passed through an account. The credential creates a defined execution role. It does not prove initiation, approval, settlement, accounting treatment, personal knowledge, personal benefit or fault.

The table-shaped hole in the record

The sentence referring to “the table above” is the most consequential line in the resolution because the table is not present on the fixed archived page available for this analysis. The absence matters for a narrow reason: the table appears, from the resolution’s own wording, to have contained the delegation detail that qualified the four names. Without it, the public reader cannot reconstruct the operating rule.

Several architectures are possible in principle, and none can be selected from the public text. A mandate might permit each signatory to act alone. It might require any two. It might require one executive and one finance or Board-facing officer. It might set one combination for routine payments and another for higher values. It might distinguish paper instruments from electronic instructions, payroll from supplier payments, transfers between company accounts from external payments, or ordinary transactions from exceptional ones. It might also limit authority by account, currency, banking product or duration.

These examples show why a control matrix matters; they are not claims about what AFRINIC’s missing table said.

That distinction between possible design and established fact must remain bright. It would be wrong to infer a two-signature rule simply because four people were named. It would be equally wrong to infer that each person could act alone because no joint rule appears in the surviving text. Silence cannot choose between mutually incompatible control arrangements.

The missing table also affects interpretation of the word “signatories.” In ordinary corporate practice, that label can cover different powers depending on the mandate: signing a cheque, approving an electronic transfer, instructing a change to account services, acknowledging bank documents, or joining another authorised person for a defined action. The public record does not specify which banking services were covered. It provides no account number, bank, branch, currency, product or transaction-class scope—and those sensitive details should not be guessed or demanded for public release.

Nor does the page establish the mandate’s lifecycle. A resolution date and a bank effective date are different facts. A board may approve a change before documents reach the bank. The bank may require forms, identity checks, specimen signatures or platform configuration. An earlier mandate may remain operative until the replacement is accepted. A later role change may require revocation even if the original resolution contains no expiry. The evidence package does not establish an earlier slate, the precise sense in which the October decision replaced it, or the dates on which any access became active, expired or was withdrawn.

The right conclusion is therefore carefully limited. AFRINIC publicly named four signatories and said a delegation table governed them. The surviving public page does not reproduce that table. This leaves the most important public design question unresolved: which permitted combination could instruct which action, under which limit, during which effective period? It does not establish that there was no private table. It does not establish that a bank lacked a complete mandate. It does not establish that an improper instruction occurred.

A bank instruction is one link, not the whole chain

The governance value of the resolution becomes clearer when the treasury process is separated into its distinct stages. Treating every stage as “approval” blurs responsibility. Treating every downstream record as proof of upstream authority allows one document to stand in for another. A serious control account instead asks what evidence exists at each link and whether the links can be joined without invention.

The first link is the underlying obligation. Before anyone tells a bank to move money, something must establish why the company should pay. That object might be an invoice supported by a procurement and delivery record, a payroll file arising from employment obligations, a tax assessment, a refund approved under a policy, or another corporate commitment. Resolution 201110.130 says nothing about any particular obligation. It names people at the bank-execution edge; it does not originate contracts or expenses.

The second link is initiation. Someone prepares the instruction, identifies the beneficiary, enters the amount and attaches or references the supporting file. Initiation can be separated from approval so that the person who constructs a payment cannot alone validate it. The public resolution does not identify initiators or show whether a named signatory could also initiate. That is an unknown, not evidence of concentration.

The third link is approval of the obligation and payment. Here the adjacent Resolution 201110.131 supplies a bounded rule: the same four people may act in accordance with AFRINIC’s authorisation policy, except that an external transaction above USD 100,000 or equivalent requires Board approval. This is evidence of a high-value escalation threshold. It is not a complete approval matrix. The record does not specify the lower tiers, approvers for particular transaction classes, exception handling or how multiple related instructions should be aggregated.

The threshold must be read exactly. “Above USD 100,000” describes the point at which Board approval was expressly required for an external transaction. It does not say that transactions at or below that amount needed no approval. It does not say that they could be single-signed. It does not turn the amount into a bank’s mandate limit. It does not prove that the Board’s approval, where required, substituted for a permitted signatory combination. Corporate approval and bank execution may be connected, but they answer different questions.

The fourth link is execution. One or more credential-holders instruct the bank in the manner the operative mandate permits. This is the link to which Resolution 201110.130 speaks most directly. It tells us the pool of four named people. Because the delegation table is unavailable, it does not tell us which member or combination of that pool could execute a given action.

The fifth link is bank acceptance. The bank must recognise the instruction as consistent with the mandate it holds. Resolution 201110.132 directed AFRINIC’s bankers to be furnished with relevant policies and resolutions, including that resolution. The direction is a sensible control component because internal authority is operationally useless if the bank’s records are different. Still, the public direction is not evidence that any named bank received the package, accepted it, configured a rule or rejected an instruction. No bank is identified in the sealed record.

The sixth link is settlement. Even an accepted instruction is not proof that money moved as requested. Settlement evidence would identify, within protected records, whether the amount, beneficiary and date matched the approved instruction. No transaction is within the subject record, so no settlement finding can be made.

The seventh link is accounting and reconciliation. Finance records the movement in the appropriate ledger, matches it to bank evidence and checks it against the supporting obligation. These acts matter, but they cannot retroactively create authority that was absent upstream. An entry in the accounts can show that a movement was recognised. It cannot by itself prove who was permitted to incur the obligation or instruct the bank. Conversely, a properly authorised bank instruction does not establish that it was later classified and reconciled correctly.

The eighth link is lifecycle review. Mandates should change when roles change, credentials are compromised, banking arrangements change or a scheduled review identifies a need. A robust record joins each person and office to an effective interval, records the bank’s acknowledgement, and shows when the authority was superseded or revoked. The 2011 public page does not supply that lifecycle. It therefore cannot support a claim that any of the four remained authorised after the period evidenced by the decision.

This chain shows why a signatory resolution is both important and incomplete. It provides one authoritative corporate object for the execution stage. It should be joined to the upstream obligation and approval, the bank’s operational mandate, the settlement and the later reconciliation. Aggregate accounts cannot substitute for those joins, because totals and classifications do not identify the authority path of an individual instruction. Equally, a missing public join does not prove a missing private record.

The bounded October control environment

Resolution 201110.130 should be the centre of the analysis. Three adjacent October decisions help describe its environment, but none should be turned into a separate controversy or used to make the signatory resolution say more than it does.

Resolution 201110.131 repeated the four people and placed their action under AFRINIC’s authorisation policy. It added the exception requiring Board approval for external transactions above USD 100,000 or equivalent. This tells us that the signatory slate was not published in complete isolation: the Board represented it as operating within a policy framework, with a specific high-value escalation. Yet the public text does not reproduce the policy. The threshold does not disclose signing combinations, and a Board-approved obligation would still need to pass through whatever execution rule the bank mandate required.

The high-value rule also raises a general design question about aggregation. A threshold protects only what its definitions capture. A competent policy should say whether related payments are combined, how staged invoices are treated, what period applies for aggregation, who may approve an exception, and how an exception is reviewed afterward. Otherwise, a series of smaller instructions could fall outside the expressed high-value trigger even though their economic substance is related. Nothing in the sealed record proves that AFRINIC split transactions or avoided the rule.

Aggregation is a test of the published design gap, not an allegation about conduct.

Resolution 201110.132 addressed the communication interface with the bank. It directed that relevant policies and resolutions be furnished to AFRINIC’s bankers. That is a necessary operational instinct: a private company cannot rely on an internal document if the bank’s mandate has not been aligned with it. The documentary chain would be stronger with transmission evidence, bank receipt, acceptance, activation date, configuration detail and later revocation. None appears in the public package.

Resolution 201110.133 established a Finance and Audit Committee for one year from 12 October 2011 and named Board members Elkins, Nyongesa, Traore and Seeburn. Its existence adds an oversight object to the sequence. Krishna Seeburn’s appearance both as a named signatory and as a committee member is a matter of public role description, but the page supplies no transaction or committee record from which to infer review, conflict or conduct. The committee’s formation does not prove that it examined a mandate, sampled an instruction or reconciled a payment.

Taken together, the sequence separates four formal objects: the signatory slate, an authorisation policy with a high-value Board exception, a direction to notify bankers, and a Finance and Audit Committee. Separation on paper is useful because it prevents one sentence from pretending to do everything. It also creates four places where evidence can be tested. But four resolutions do not prove that the objects were integrated in operation. A policy may exist without being configured at the bank; a bank mandate may operate without being mapped to transaction approval; a committee may exist without reviewing either.

Integration requires joined records.

The archived Board page associates this October sequence with a teleconference whose final resolution records adjournment at 17:20 UTC on 12 October 2011. That provides meeting context, not a complete adoption record for Resolution 201110.130. The surviving entry does not specify the exact adoption time, attendance, quorum, mover, seconder, vote count, abstentions or conflicts for the signatory resolution. Those are evidentiary limits, not reasons to invent procedural invalidity.

The strongest defence and the strongest critique

Fair institutional analysis should state the best case for the visible design before identifying what it cannot prove. The defence here is concrete. AFRINIC named people rather than hiding behind an undefined class of officers. It attached offices to the names. The selected roles reached across executive leadership, finance, administration and a Mauritius-based Board representative. The Board connected the slate to an authorisation policy, imposed a Board-approval threshold above USD 100,000 for external transactions, told bankers to receive relevant instruments, and created a Finance and Audit Committee for a defined one-year period.

Those are recognisable components of treasury control. Publicly naming the signatory slate improves accountability because a later reviewer can ask whether the bank mandate matched the Board’s instrument and whether a transaction was executed by someone within the approved pool. A high-value escalation can constrain exposure. Bank notification can align corporate and banking records. A committee can provide an oversight venue. The sequence therefore supports an inference of plausible control design intent.

But design intent is not implementation, and implementation is not effectiveness. The public record does not show the delegation table that would make the execution rule testable. It supplies no bank acknowledgement. It gives no effective or revocation dates. It identifies no account class, transaction class or permitted combination. It provides no transaction ledger, exception log, rejected-instruction log, access-token inventory, specimen-signature custody record, periodic mandate review or reconciliation sample.

The critique should be just as concrete: by omitting the referenced table, the surviving page withholds the rule most necessary to interpret the four names. Readers cannot verify whether the intended architecture actually separated execution or merely distributed potential credentials. They cannot see which conditions applied at which value or for which period. And because the bank-acceptance and lifecycle layers are also absent, they cannot follow the corporate act into operational effect.

That critique does not require alleging fraud, theft, illegality or even bad administration. Public transparency can be incomplete while private controls are sound. A historical web page can lose a table while the original resolution, bank forms and internal files remain intact. The evidence supports a demand for a better public control record and protected independent verification. It supports no claim that any person used the mandate.

What the record cannot carry

The evidentiary boundary is part of the finding, not a footnote. No sealed source proves that Adiel Ayodele Akplogan, Patrisse Deesse, Krishna Seeburn or Christian Fanchette signed a particular instruction. None proves that any of them initiated or approved an underlying obligation, received money, held a conflict, benefited personally, exceeded authority or remained authorised after October 2011. None proves a bank rejection, an invalid mandate, a loss, theft or unauthorised payment.

The record also does not identify any AFRINIC account, bank, branch, currency, balance, password, token, routing detail, owner or beneficial interest. It should not be stretched to supply those facts. Publishing sensitive access information would weaken security rather than improve accountability.

The official page has a narrower evidentiary role. It proves what AFRINIC published as a Board resolution sequence. It is not self-authenticating proof that every stated act was validly adopted, fully implemented, representative or effective. Corporate records should be checked against the authority and evidence appropriate to their function. A resolution can establish a documented internal act. A bank acknowledgement establishes a different operational fact. A transaction record establishes another. An accounting entry establishes another still.

The inverse also holds. Absence from the public page does not establish absence from AFRINIC’s private files. The missing table may have existed elsewhere. The bank may have held detailed instructions. Internal reviewers may have tested them. None of that is proved by this package, but neither may it be denied merely because it is not online. Institutional analysis earns trust by preserving unknowns on both sides.

Private treasury control does not create public authority

The bank mandate concerns AFRINIC’s own private corporate administration. AFRINIC performs a narrow bookkeeping and coordination service for Internet number-resource records. Its legitimate usefulness lies in accurate, unique and contactable records, fraud-resistant maintenance, transfer recording, security assertions and operational continuity. It is not a state, legislature, regulator, police force, punishment authority, confiscatory authority or adjudicator.

That distinction matters here because internal competence is easily converted into an external legitimacy story it cannot bear. A well-designed bank mandate could show that the company administered its own money responsibly. It could not turn an internal policy into law for network operators. It could not create sovereign authority over number resources. It could not make a registry entry the source of the underlying rights or operational reality the entry describes.

A weaker mandate would not produce the opposite transfer of power. Gaps in a private company’s controls can expose agency risk and justify demands for accountability, but they do not give critics, members, policy participants or supporters a confiscatory or adjudicatory power either. The remedy remains within the proper institutional frame: require the company to make its mandate-control record testable, protect sensitive evidence, and preserve the technical ledger’s continuity.

This is why treasury accountability and registry authority must be kept analytically separate. Operators and productive businesses can bear real continuity and financial costs when a registry’s internal governance fails. That exposure makes good controls important. It does not make the registry sovereign. The ledger records; the company coordinates; neither function creates the reality it records or a throne from which to command it.

What a complete historical account would join

A complete account of the 2011 control would not begin by publishing raw bank files. It would begin by assigning every evidentiary object a distinct role. The Board resolution would identify the corporate decision and approved signatory pool. The operative delegation policy would identify roles, combinations, transaction classes and amount tiers. Bank correspondence would establish receipt and activation. Access records would identify credential provisioning and revocation. Transaction evidence would connect a specific obligation, approval, instruction, settlement and reconciliation.

Review records would show whether exceptions and mandate changes were tested.

The join should be possible transaction by transaction inside a protected review environment even if only bounded findings are made public. For a sampled payment, an independent reviewer should be able to answer: What created the obligation? Who initiated the instruction? Who approved it under which policy version and amount tier? Which authorised combination executed it? Did the bank accept and settle exactly that instruction? How was the result recorded and reconciled? Were exceptions documented? Were all credentials current on that date?

Each answer needs its own evidence. A later account balance cannot identify who had upstream authority. A Board approval cannot prove bank execution. A bank statement cannot establish the corporate basis for the obligation. A reconciled ledger entry cannot manufacture missing authorship. An audit or committee review can test evidence and report findings, but it does not become a court judgment or replace the instruments being tested.

For the historical 2011 slate, the public record closes only a limited portion of this chain. It identifies four people and their offices, references a delegation table, shows a high-value Board threshold in adjacent context, directs banker notification and establishes an oversight committee. It does not close the table, bank, lifecycle or transaction layers. That is a precise governance result: more than an empty assurance, less than an auditable mandate history.