n.V5-7.05 | MEDIATION
01| Projects rarely fail at once. They fail through friction that nobody resolved: a disagreement inside a team that hardened into two camps, a professional who took over the work instead of supporting it, a buyer whose expectations drifted from what was agreed, a decision at CITI level that nobody explained. Each of these is small on the day it appears and expensive three weeks later. Mediation exists because the cost of unresolved friction is not the argument. It is the project.
02| This has to be stated as a design commitment rather than a good intention, because every organization believes it handles conflict well and most handle it by waiting. The default response to friction is to hope it settles, which works often enough to be dangerous. When it does not settle, the delay has already done the damage: positions have been taken publicly, people have chosen sides, and what was a disagreement about a decision has become a disagreement about people.
03| Five kinds of friction are predictable, so they get designed for. Inside a Kikundi team — workload, credit, direction, and the ordinary difficulty of working with people you did not choose. Between a team and a professional from Pro.Crowd or the ecosystem, which is the most dangerous category, because the expert usually knows more and can therefore take control without anyone deciding that they should. Between a team and a buyer, over scope, delivery or whether what was promised was delivered. Between a team and CITI leadership, over resources, priority or a decision that landed without explanation. And within a project that is simply going wrong mid-flight, where the argument is a symptom and the project needs turning around rather than adjudicating.
04| The principles are few and they are strict. Mediation is not judgment. The mediator has no authority to rule and no stake in the outcome; the job is to get the parties to a workable answer, not to declare a winner. The moment mediation becomes a court, people start preparing cases instead of solving problems. Speed beats correctness. A decent resolution this week is worth more than an excellent one next month, because the project is running the whole time. The process is private, the outcome is not. People will not speak honestly if the conversation is published, but a resolution nobody can see is indistinguishable from a favour. So the discussion stays closed and the decision, with its reasoning, goes on the record. Escalation exists and is published in advance. Anyone can see, before they need it, where a problem goes if mediation fails. An escalation path that has to be discovered by asking around is a path only the well-connected have.
05| The professional-takeover case deserves its own rule, because it is where the whole model is most quietly at risk. When a Maveriq team brings in specialist expertise, the natural gradient runs toward the specialist: they are more experienced, more fluent, and usually right about the technical question. If that gradient runs unchecked, the team becomes staff on its own project and we have rebuilt the top-down model inside our own structure, with better branding. Mediation in these cases starts from a fixed position — the team owns the decision, the professional owns the advice — and works out how to keep both true. A specialist who cannot work that way is a poor fit for this platform regardless of how good the work is.
06| Support is the other half of the same function, and it is the cheaper half. Most friction is not conflict; it is overload. Maveriq overload is one of our largest risk factors, because the people carrying the system are also carrying unstable lives, and the failure mode is not a dispute but a quiet disappearance. So support runs on two tracks. Practical load is lifted off wherever possible — logistics, supply and delivery handled by the one-stop shop rather than by a Maveriq at nine at night. And life pressure gets an assistance programme, on the same premise as the readiness work that precedes all training: a person whose life is in chaos cannot perform, and pretending otherwise does not make them perform. Neither track is charity. Both protect the engine.
07| Two boundaries keep this from becoming its own bureaucracy. Mediation does not decide what a project should be — that authority sits with the team. And mediation is not a route around the market: a team that loses a buyer through poor work is not entitled to a mediated second chance. The line is between friction that blocks good work and consequences that good work would have avoided.
08| Finally, mediation is data. Every dispute is a signal about a rule that is unclear, a boundary that is wrong, or an incentive pointing the wrong way. Individually they look like personality problems; in aggregate they are almost always design problems. So the pattern of disputes — what kind, between whom, how often, resolved how — feeds directly into the rule review described in the next chapter. A platform that mediates well and learns nothing from it will mediate the same argument forever.
