n.V5-7.06 | FAILURE MODES

01| IkoCiti is not a neutral tool, and any claim that it is would be dishonest. Social networks have spent two decades insisting that users create the content while their algorithms quietly determine what anyone sees. We are in the same position and we should say so plainly. Our fingerprints are on everything: selection criteria decide who participates, compensation structures shape behaviour, quality standards enable or block action, governance decides who decides, AI design touches every interaction, and the architecture bounds what is possible at all. So the excuse is not available to us. We cannot say the tool worked and the users failed. Failure modes here are not external events. They are built in — and naming them honestly is the only way to design against them.

02| The first category is rules that are wrong, usually by being too tight or too loose. Selection rules exclude capable people when tight and admit people who cannot deliver when loose — and the same applies to which barrios get a CITI, which buyers are accepted, and which professionals plug in. Quality standards paralyze the system when too strict and produce nothing; when too loose they let through the projects that collapse our reputation. Track-record metrics get gamed when they are rigid; documentation requirements either kill energy with bureaucracy or fail to capture anything worth keeping. Compensation is two-sided at every point: a base too high creates dependency and too low loses people; training pay too high crowds out the actual work and too low means nobody trains; maintenance pay too high crowds out new projects and too low lets impact evaporate anyway. Process rules drown the system in planning or let fragmentation creep back through under-coordination. AI rules produce blind dependency on machine judgment, or fail to give Maveriqs the leverage AI is supposed to provide. Governance rules centralize into bottlenecks or fragment into local capture. Ecosystem rules either recreate the top-down model or exclude expertise we genuinely need.

03| The second category is rules we do not have and should. Missing quality control produces shoddy work and reputational collapse. Missing coordination enforcement lets fragmentation return. Missing follow-up requirements let impact evaporate. Missing predator protection lets bad actors work the barrio through us. Missing limits on professional takeover rebuild the old model inside the new one. Missing data privacy rules invite exploitation and destroy trust. Missing exit criteria for failed CITIs create zombie platforms draining resources without producing anything. Missing conflict-of-interest rules invite capture. Missing escalation paths let problems fester. Missing Maveriq protections produce burnout. Missing limits on AI autonomy let bad decisions scale. Missing term limits invite entrenchment. Missing whistleblower protection keeps problems hidden. Missing appeal processes let arbitrary decisions stand. Missing sunset clauses keep dead rules alive.

04| The third category is design choices, which are the most expensive because everything downstream inherits them. Architecture: a single platform with no redundancy is a single point of failure; centralized data makes one breach catastrophic; a complex interface does not land in the barrio; mobile-only excludes people without phones; internet-dependent fails where connectivity does. Organization: too flat and accountability disappears; too steep and bottlenecks appear; division-based produces silos; matrix produces confusion. Funding: a competitive exchange can produce zero-sum thinking where we wanted collaboration; subscriptions break the link to performance; fee-for-service narrows everything to transactions; grant dependence puts us exactly where the NGOs are. Maveriq model: part-time only weakens commitment; full-time only excludes the working poor; individual focus erodes team accountability; team-only fails to reward individual excellence. Incentives: output-based pay invites gaming; input-based pay severs the performance link; individual bonuses produce competition where we wanted collaboration; team bonuses produce free riders.

05| Every item above is a real trade-off, not a solvable problem. That is the point of listing them. Anyone who reads this list and concludes that we should simply choose the right setting has misunderstood: there is no setting that is right in all conditions, which is why the settings have to be measured and revised rather than decided once and defended.

06| Accountability is assigned so that the list can be acted on. Wrong or missing rules are IkoCiti Global's responsibility — rule-setting sits with us. Failed platform design is also Global. Local execution failing despite working infrastructure is CITI leadership. An individual Maveriq failing sits with the Maveriq and the support system around them. External shocks overwhelming a CITI belong to nobody, but are tracked anyway, because unexplained variance is how organizations start lying to themselves. And if nobody wants the tool at all, that is a value proposition problem and it is entirely ours.

07| Some of this can be tested early and some cannot, and being clear about which is which prevents false confidence. Testable within months: whether selection criteria produce the right people, whether the compensation level is sustainable, whether quality standards are enabling work or blocking it, whether AI support helps or creates dependency, whether the coordination process is too heavy or too light. Testable only over years: whether self-perpetuation actually works and trained Maveriqs successfully train others, whether impact survives after the project ends, whether the talent pool compounds, whether CITI replication scales. The second list contains our largest claims. We should be honest that we will be operating on faith for a long time before evidence arrives.

08| So the rules get reviewed on a cycle. Every rule is documented — what is it. Justified — why does it exist. Measured — is it working. And revisable — when do we change it. Annually we ask which rules caused problems this year, which absent rules allowed problems, what to add, remove or modify, and what the pattern implies about the design. The dispute record from mediation feeds straight into this, because arguments are usually rules failing in public.

09| This is the discipline that keeps us from drifting into the posture every mature institution eventually adopts: blaming the people it exists to serve for outcomes its own design produced. The system has rules. The rules have consequences. We own them.