Designed by insiders, inflicted on everyone else
An email from a trusted sender has gone to Junk again.
I have rescued it before. More than once. Each time Outlook has smiled sweetly and assured me that future emails from this address will not be treated as junk. Yet here it is again, sitting in Junk like a repeat offender with diplomatic immunity.
After some investigation, it appears the culprit might be Outlook. Or Hostinger. Or both. Or neither, depending on which layer of the system made the decision before the other layer displayed it.
This isn’t really about email.
It’s about a recurring feature of modern life: systems designed by insiders and inflicted on everyone else.
We meet these systems everywhere. In software. In banking. In website design. In government forms. In public services. In church administration. In the bright, smooth, confident language of organisations that tell us how simple everything is, while quietly leaving us to puzzle out which department, platform, account, password, setting, portal, filter, policy, or hidden rule has decided to frustrate us this time.
The pattern is always much the same.
The consumer has a simple human task:
I want this email to arrive in my inbox.
I want to renew this policy.
I want to publish this page.
I want to speak to someone who can sort this out.
I want to know why this form has been rejected.
I want to understand what I’m supposed to do next.
The system does not answer that human task. It answers from inside its own machinery:
Please check your account settings.
Please log into the portal.
Please contact the other department.
Please confirm your identity again.
Please use the new interface.
Please use the old interface.
Please note that this feature is not available in the new interface.
Please note that the old interface will shortly be retired.
Please do not reply to this email.
And so the consumer’s task fails while every department’s process succeeds.
That, I think, is the heart of it.
No one has done anything obviously wrong. The email provider might have run its spam filter. The desktop client might have displayed the folder correctly. The help page might have described one version of the software. The bank might have followed its security protocol. The church office might have used the correct form. The government department might have met its response target.
And yet the person at the receiving end is still stuck.
Worse than stuck, they are made to feel faintly foolish. As though the failure must be theirs. As though a competent adult should naturally understand the difference between an alias, an account, a mailbox, a sender, a domain, an IMAP folder, a Microsoft identity, a webmail setting, and a junk-mail rule.
This is one of the quiet cruelties of badly designed systems. They don’t merely waste time. They transfer blame.
A badly designed system whispers, “You should have known where to click.”
A badly designed system mutters, “You should have read the instructions.”
A badly designed system implies, “Other people manage this. Why can’t you?”
But perhaps other people don’t manage it. Perhaps they merely give up. Perhaps they find a workaround. Perhaps they phone a younger relative. Perhaps they pay someone. Perhaps they abandon the task altogether. Perhaps they tell themselves they are too old, too stupid, too slow, too out of touch.
And the institution records none of that as failure:
The call was answered.
The form was available.
The email was sent.
The guidance was published.
The portal was functioning.
The customer was advised to check the relevant settings.
Process successful. Human task unresolved.
This isn’t only a technical problem. It’s a moral one.
Systems should serve people. That sounds almost too obvious to say. Yet again and again we encounter systems that appear to have been designed primarily to serve the organisation that owns them. They reflect the organisation’s departments, anxieties, liabilities, histories, jargon, internal boundaries, legacy software, and budget compromises. The consumer is then expected to learn the map of that private kingdom in order to complete even a simple task.
And if the consumer fails, that too becomes part of the system. Another support ticket. Another complaint. Another frequently asked question. Another chatbot conversation in which artificial intelligence is invited to explain why the unintelligent design has confused yet another human being.
This is where the current enthusiasm for AI can either help or make things worse.
Used badly, AI will simply become a velvet curtain hung over a locked door. It will apologise fluently, explain vaguely, and keep the consumer moving in circles. It will become a more articulate version of “computer says no.”
Used properly, AI could do something far more valuable. It could interrogate the system itself. It could say:
This email was put in Junk by Hostinger, not Outlook.
This claim is delayed because one document is missing.
This page will not publish because the template is assigned incorrectly.
This form has been rejected because two departments are asking for contradictory information.
This is the person or team responsible for correcting it.
That would be useful. Not artificial intelligence sprinkled over bad design like parsley on cold soup, but intelligence used to expose bad design and remove it.
The demand should be simple.
When a system blocks, moves, rejects, delays, downgrades, or refuses something, it should tell the affected person three things:
Where the decision was made.
Why the decision was made.
Who or what can correct it.
That should be a basic standard of decent administration.
It applies to email. It applies to banking. It applies to software. It applies to public services. It applies to the Church. It applies to politics.
Indeed, politics might be where the same pattern does most damage. Governments and public bodies can follow every internal procedure and still leave citizens feeling unheard, baffled, and powerless. A minister announces a policy. A department writes guidance. A contractor builds a portal. A local office processes a form. A helpline reads from a script. Each part might claim to have done its job.
But the citizen’s task has failed.
And then politicians wonder why trust drains away.
Trust isn’t rebuilt by slogans about transparency, customer care, digital transformation, or putting people first. Trust is rebuilt when systems tell the truth, accept responsibility, and help people complete the task they actually came to complete.
The problem isn’t that modern life is complicated. Some of it inevitably is. The problem is that institutions too often pass their own complexity on to the people least able to untangle it.
We don’t need to be flattered. We don’t need cheerful pop-ups, glossy dashboards, or endless invitations to rate our experience. We need systems that say plainly what has happened, why it has happened, and what can be done about it.
Until then, the consumer’s task will continue to fail while every department’s process succeeds.
And institutions will continue to wonder why people are angry.
Fix the system, not the user.
Leave a Reply