A private charter desk handles information that clients expect to stay private: names, travel patterns, preferences, sometimes security arrangements. It also relies on a crowd of people who need to touch it: coordinators, crew schedulers, ground handlers, a software supplier, a freelancer who fixes the website. If you have ever wondered whether too many people can open too many doors, you are asking an AccessOps question. Below are five symptoms with a diagnosis and a fix for each. No fluff, and no promise that tooling alone will save you.
Terminology first. The Ops Atlas describes AccessOps in its reference entry on requested and privileged access as the operational side of granting rights beyond a person’s ordinary ones. It carries no official standing, a point the Atlas makes without hedging. The practices inside, such as elevation on request, emergency accounts and session recording, are long-established.
Symptom 1: Nobody can say who has the booking system’s admin rights
Diagnosis. There is no inventory. Rights were added over years, one favour at a time. The Atlas’s maturity ladder starts exactly here, at “ad hoc”.
Fix. Spend a day listing every administrator on the booking platform, the email system and the payments account. Put the most sensitive in a top group and the rest in a second. This mirrors the Atlas’s first thirty-day step, which covers the top two tiers and one emergency drill. Anyone who cannot be explained loses access until they ask again.
Symptom 2: Contractors keep their logins long after the job
Diagnosis. Grants have no end date, so they live until someone remembers.
Fix. Every contractor grant gets an expiry written at the time it is made. The Atlas suggests a default of three to six months for ordinary requests and far shorter windows for the highest-risk systems: one to two hours for the top tier, two to four and four to eight below it. A website freelancer fixing a plugin should not have the same window as someone who can export all client records. Then check that the session actually ended, as the Atlas points out that a login already in progress can outlive the permission that created it.
Symptom 3: Approvals are a rubber stamp
Diagnosis. A manager approves every request in seconds because the reasons all say “needed for work”. The Atlas names blanket approvals as a pattern that breeds generic justifications.
Fix. Require a specific reason, such as “re-sending a manifest for the Thursday positioning flight”, and record five items per grant: holder, scope, reason, end time and approver. For routine, low-tier requests a rule is fine, for example “coordinator on the live roster”. For anything touching passenger data in bulk, ask a second person. Before you automate any of this, note that the entry treats an automatic approver as part of the top tier itself, and that it has found no documented safe way to hand top-tier approval to a language model.
Symptom 4: The emergency login is either missing or used daily
Diagnosis. Either there is no way to recover if the approval process fails, or the emergency account has become the normal route because it is quicker.
Fix. Keep two emergency accounts, as vendor guidance cited by the Atlas recommends, stored away from the daily workflow and protected with phishing-resistant sign-in. Test them every ninety days or sooner, and have every use trigger an alarm followed by a next-business-day review. One cloud provider’s documentation quoted in the entry says that handling such an account casually can make it a backdoor. A related trap: if your approval rule has an empty list of approvers, you can lock yourself out of your own tenant, so test changes before applying them.
Symptom 5: Third parties can see everything
Diagnosis. A supplier’s support team can enter your systems whenever it wishes.
Fix. Ask for the pattern one large cloud provider offers in its lockbox feature, as the Atlas summarises: staff access capped at about four hours, a request that expires after twelve, and your explicit approval. If a supplier cannot offer something similar, ask what logging you will get. The entry also recalls a 2022 compromise of an outsourced support workstation: 366 customers had to be considered at first, and the review narrowed the real impact to two tenants. Good records are what make that kind of narrowing possible.
Measure the repair
Fixes you cannot measure fade. The entry suggests six simple gauges. Count the people with permanent rights in the top two tiers. Work out how much privileged work happens through on-request elevation. Time how long approvals take, looking at a typical request and at the slowest tenth. Compare granted minutes against minutes actually used. Note what fraction of sessions are recorded. And track the lag before any emergency login is examined. In its worked month, the permanent count fell from 34 to 9, and the emergency uses left only 7 items to explain. A charter desk will have much smaller numbers, but the direction counts.
Speed matters too. If obtaining access takes half a day, coordinators will share passwords to meet a departure slot, and you will have recreated the original problem. Aim for a short median and watch the slow end.
A sample week after the fixes
It helps to picture what the end state looks like in daily use. On Monday a coordinator needs to correct a booking in bulk. She raises a request naming the charter reference, a rule approves it because she is on the live roster, and the elevated rights lapse that afternoon. On Wednesday the website freelancer asks for plugin access for two days; the owner approves, notes the end date and sees it disappear on Friday without a reminder. On Friday afternoon the desk lead spends ten minutes reviewing the one emergency login use from earlier in the week, which turns out to be a scheduled test. Nothing dramatic happened, and that is the intended result.
Where to go from here
After the diagnosis, work in order: inventory, time limits, reasons and approvals, then recording and daily reconciliation for the top tier. Pair the work with the operational planning you already do. Our passenger-side list for planning a rotorcraft charter is built around thinking ahead of departure, and access deserves the same lead time. When you review the client-facing side, this piece on what luxury travellers expect from providers is a reminder that discretion is part of the product.
Charter is a trust business. Knowing who can open the passenger file, for how long and why is as much part of the service as the aircraft on the ramp.

