I want you to picture, just for a second that you’re in 1700s Britain, arriving home after dark.
Before you can even see the hallway, you feel your way around the ledge beside the door for the tinderbox, or perhaps you leave it in a dry place outside for moments like this. You strike a piece of flint across the steel loop into the charcloth inside the box until you see an ember catch. Blowing on the cloth until it catches alight, you can now light the last few centimetres of candle wick, you remembered to put back next to the box.
If you want to go anywhere in the house, you’ll need to carefully shield that candle and carry it with you until you find the next placed one to light, and repeat. You are constantly managing it. A sudden draft could just blow it out, or worse. You could accidentally set fire to your curtains and burn the whole house down. You are entirely responsible for managing the mechanism of your own lighting.
It’s an exhausting way to live. In 100 years from now glass will become industrialised making hurricane shields (glass lanterns) more commonplace. Another 50-60 years and if you’re part of the upper-classes, you may be able to afford kerosene for oil-lamps that burn longer and brighter.
For now, you settle with the tallow based candles, rendered animal fats which are incredible cheap to buy, or make yourself and constantly serving the light source rather than the light source serving you. Spending more time managing the tool, than actually benefiting from the outcome.
Along with Kerosene, luxury burning materials such as whale oil has come and gone, their entire industrial complexes with it. With each new medium the approach to solving the problem - lighting a room will just be iterations on the same. Burn a fuel to generate heat, which generates light.
The electric grid didn’t give humanity a better, safer candle.
Since the turn of this century, the prevailing wisdom was basically that taking a bureaucratic service and just throwing it onto a website was the ultimate solution. The holy grail of modernisation. Holding onto this wisdom has ignored what we learnt from Silicon Valley and missed the entire point of the digital revolution.
We fundamentally misunderstood what the web was actually doing. The candle analogy perfectly illustrates our current deeply flawed relationship with digital services.
If we connect this to the bigger picture, think about what actually happened when electricity was introduced to the home. The electric grid didn’t just give humanity a better, safer candle. It wasn’t just a candle 2.0.
What electricity did was separate the outcome, which is the light in the room from the mechanism required to produce it. When you walk into a room today, you just flip a switch and the light comes on. You don’t need to know how the power plant 50 miles away generates megawatts. You don’t need to understand alternating current or transformers or how the copper wiring inside your walls is routed. The system absorbs all of that complexity. The burden is completely on the infrastructure, so you just get the outcome.
But here is the massive reality check when it comes to digital transformation of any large institution. We haven’t actually invented electricity yet.
We’ve basically spent the last 25 years building digital candles. We took the exact same cumbersome high maintenance process as we used to do on paper and then we put them on a screen. Word processor files (Word), then PDF on a screen, instead of a printed sheet and a clipboard. We treat the website as if it is the service itself. We assume that because an organisation has a website, they have modernised. But the actual service being provided, like applying for a business permit or registering a vehicle - that existed long before the internet. It existed at a physical front desk with a clerk and a postal system to a central office. After that, it was FAX over a landline and a telephone call. The website is just the newest iteration of the same front desk part of the service.
The interface, whether it’s a heavy mahogany desk in a stuffy building, an 0800 number, or a responsive web page, is simply a mechanism for an organisation to increase its reach and reduce its operational costs. It’s an efficiency play. I Know. In the last 16 years I have worked on dozens of transformation projects where the goal, and definition of success was explicitly to make reduntant 100s of people.
This is the underlying reality for any organisation delivering high volume services. Running a centralised website on a server is infinitely cheaper than staffing a massive call center. Which, in turn, is cheaper than maintaining 50 physical offices. So the shift to digital is primarily about the organisation saving money. But changing the interface to save the organisation money does not fundamentally alter the architecture of the services being delivered to the people accessing them. The core mechanics are exactly the same.
The interface changes, but human needs rarely do
If the website is a constantly optimising interface to save money, what isn’t changing? We have mobile apps, we have cloud computing, AI chatbots. We have all the technology. Why does, in particular dealing with a government service still feel like you are navigating a bureaucratic hedge maze blindfolded?
It’s because the interface changes, but human needs rarely do. The technology evolves, but the fundamental underlying reasons why a person needs to interact with a public service organisation remains remarkably constant across generations.
Human need is a pretty broad term. Let’s ground it in reality. A major life event doesn’t care about bureaucratic boundaries. Think about a new parent. If you just had a baby, you aren’t just dealing with one single unified department, or service.
You need to register a birth, which is local government. If you’re in the UK, it is your responsibility to register your child with your GP - despite their NHS number being automatically generated by the hospital they were born in (assuming they were born in a hospital). This is another contact with potentially an entirely different NHS Trust. You’re not done. After this you might need to apply for child tax credits, which will require you to create an account and apply through HMRC.
You may need to notify your employer, who can update your records, including information about childcare, benefits packages, pensions and a host of other things. If you’re travelling internationally soon, you need to apply for a child passport for your baby which, you guessed it, requires another interaction with another central government department.
Life insurance, Wills, Medical Insurance, Pensions, Car Insurance, Home Insurance, it’s this messy mix of private and public systems.
For the first time in the web-age the majority of people access the internet via a small-screen called a phone instead of mailing in carbon copy forms around each of these organisations, but the fundamental need to establish that child’s identity and secure their future, that hasn’t changed at all. These types of need are timeless.
Consider someone who suddenly becomes severely ill. Their core needs are immediate, profound, and universally understood. They need medical treatment first and foremost. But they also need financial security if they can no longer work. And that triggers disability claims. They might need additional life support to manage the cascading effects of that illness on their living situation or their dependants.
You can even see this in the business sector. Everyone starting a new business today needs to understand their legal liability. They have to secure permissions to operate, register for taxes, get a bank account, Insurance, numerous now mandatory software applications, find premises to work from, and declare who is running the business. In general, the exact same things someone opening their first hardware store a hundred years ago needed to do.
If these needs are so universal and so unchanging, my instinct is to say, well, build a single app that handles it all. Just one magic button. But I’m guessing that fails because the friction isn’t actually in the software code. The friction is in the actual structure of the government itself.
The human API
The friction point is exactly where human reality violently collides with organisational architecture. Not one of those life events, birth, illness, starting a business, none of them fit neatly into one single government silo. A major life event inherently crosses dozens of boundaries. It’s never just one department. It requires the action of different professions governed by totally different legislative policies, funded by different budget appropriations and stored on entirely different legacy technology stacks.
Department A is legally mandated to protect your data while Department B desperately needs that data just to issue your funding. And because Department A and Department B can’t or maybe just won’t talk to each other, this service user basically becomes the human API.
The system forces you, the public, to carry the burden of figuring out the government’s internal complexity exactly at the moment you need help the most. It’s brutal. It’s like the government is essentially saying, hey, congratulations on the new baby. We know you haven’t slept in four days, but now please navigate these six different departmental websites, figure out which one handles tax credits versus health registrations and manually reenter your address and income into completely separate incompatible databases indefinitely until you cannot see the screen anymore, one-handed, on your phone, holding a wriggling newborn.
You are the one carrying the candle from room to room.
Since the core problem is that we all have to manually navigate all this internal bureaucratic complexity, the obvious solution seems to be hiding it. Sweeping it away. We hear this buzzword all the time in tech and design - seamless.
Seamless experiences is everywhere. Just sweep all the departmental infighting in incompatible databases under a digital rug and give the user a totally blank, smooth, frictionless screen. A frictionless screen sounds ideal in theory, but making things entirely invisible is actually a really dangerous trap.
In 2026 the UK Government published The roadmap for a modern digital government which basically argues for public services that are joined up around the actual needs of people where the government, not the citizen, carries the burden of its own complexity. But hold on.
Seamful, not seamless design
If the government is carrying the burden, shouldn’t it be completely invisible to me? If we intentionally leave the seam showing and tell the citizen exactly which specific sub-department is currently reviewing their medical claim, doesn’t that just pass the anxiety back to the user? How is seeing the internal plumbing better than a seamless, invisible process where I just don’t have to worry about the bureaucracy at all?
Because complete invisibility destroys trust. A few articles ago I mentioned Platform Land by Richard Pope which advocates for a concept called Seamful Design - Seamful. Not seamless.
Seamful doesn’t mean exposing the messy internal wiring or the conflicting database queries. That means intentionally making specific boundaries visible when that visibility is useful and necessary for the citizen to understand what is happening to them.
The best anchor for this is tracking a shipped package. When I order something expensive online, I absolutely do not want a completely seamless and invisible experience where I just click buy and then sit in the dark for a week hoping a box eventually materialises on my porch.
That would be completely seamless, but it would be terrifying. I would constantly wonder if my order went through, if it was lost, or if it was stolen. I would have zero agency and zero trust in the system’s ability to deliver.
But on the flip side, I also don’t want to see every single useless seam. I don’t want an app notification showing me the license plate of every truck the package gets loaded onto or the lunch schedule of the warehouse workers in Reading.
What I want are the useful seams. I want to know exactly which courier company currently has custody of my box, what day it’s expected to arrive, and the name of the person who signed for it.
The package tracking analogy maps flawlessly to public services. The real challenge for a modern digital government isn’t about hiding every single seam through a slick user interface. It’s engineering the back end so that you can decide which seams the public actually needs to see to feel secure and for you to be trusted by them.
What does a useful seam look like in a digital government context?
A useful seam is knowing exactly which specific organisation is currently holding your sensitive medical information. It’s knowing when appropriate the identity or the role of the official who made a decision regarding your tax status - or whether a person decided at all, or if it was based on algoramith determination through an AI workflow. If a decision is made about your disability benefits by an automated algorithm, a useful seam is exposing that fact and providing a clear path to appeal to an actual human. Because otherwise, you just feel like you’re fighting a ghost.
Those are seams you must see to ensure accountability and trust. The internal squabbles over departmental funding, the incompatible mainframes, the legacy legislation that slows down data transfers. That’s just noise.
Those are the useless seams, ready to be designed out of view.
This really exposes the ultimate limitation of just slapping a new website on an old problem. A beautiful user interface cannot reconcile conflicting government policies. A website can guide you through a form using a really nice font, clear language, smooth loading animations, but a webpage cannot legally compel His Majesty’s Revenue & Customs to accept a digital signature that the Driver And Vehicle Licensing Agency requires. A website cannot create shared accountability between two departments that have completely separate budgetary mandates.
Organisations obsess over the digital interface. They are zooming in on one single tree in a huge forest.
When you stare at the tree you have lost sight of the forest
We isolate this one tree, the website, and we study its shape, we test its branches, we make it look incredibly healthy and user-friendly, but we completely ignore the forest it lives in. A public service is a deeply complex forest. Its actual health depends on the unseen interactions of the soil, the water table, the climate, the underground root systems. In government terms, those roots and soil are the data sharing agreements, the internal security protocols, the legislative mandates, and the procurement processes. You can have every single one of those back-end elements working exactly as intended within their own isolated silos. The legal team is strictly following the law. The security team is locking down the data. The finance team is balancing their specific budget, doing their jobs perfectly, but when all those optimised silos combine, they create an experience for the person accessing the service that is totally fragmented, repetitive, and infuriating.
The forest is dying, even if that one tree looks great.
Historically, a user experience designer is invited into this massive complex ecosystem, often entirely too late in the game, and at far too narrow a point in the process. They are handed a rigid set of constraints and basically told to just make the best of it. They are essentially brought into a room and told, okay, we have already drafted this incredibly restrictive, confusing policy. We have already locked in a funding model that pits two departments against each other. Oh, and we bought a proprietary database system five years ago that doesn’t integrate with anything else. Now, can you please design a web form that makes this easy for the user?
Making the form easier to read and accessible, sure, it’s important, but it leaves the actual service, the engine itself, completely undesigned because the policy is the service. The funding model is the service. This raises an important question about the future of work in this space. This realisation forces a radical expansion of what a public service designer actually does. If the web page is no longer the ultimate solution, the role of the designer has to expand from tweaking user interfaces to actively orchestrating an ecosystem.
How does a designer actually orchestrate an ecosystem? How do they shift from designing a screen to designing something abstract like governance? What are the actual mechanics of that?
Firstly, it means their raw materials fundamentally change. A designer’s materials, interface designs, journey maps, testing suites, shape and form, must expand to include employee incentives, organisational relationships, and flows of information. That’s a huge shift.
Let’s look at a concrete mechanism for this. When a systemic designer uses employee incentives as a design material, they aren’t writing code. They are literally sitting down with the human resources department and agency directors to change how a frontline caseworker’s success is measured. If a caseworker’s key performance indicator, their KPI, is strictly the number of forms processed per hour, they are financially incentivised to rush. To reject incomplete applications, to just churn through people. They ignore the complex context of a persons life because the system rewards them for speed, not for actually helping the person. The designer intervenes at the KPI level working to shift that metric from forms processed to successful person outcomes or time to secure housing.
If you don’t change the internal incentive structure, the most beautifully designed web form in the world is completely useless. Because the organisation itself is still working against the person accessing their services.
We have to stop trying to build a better digital candle and start designing what electricity makes possible.
For The designer, this doesn’t mean you have to suddenly become constitutional lawyers or database architects. You don’t have to write the legislation yourself. But you do have to become experts in seeing relationships and mapping friction. You build service blueprints instead of web pages. You have to map out how a minor data privacy policy drafted in a basement office in one department creates a disastrous real-world consequence for a sick person trying to claim benefits in another city. You have to design the invisible bureaucracy.
Your job is to take the behaviour of this massive, messy, disparate system, map it out and make it totally visible to the people in power, the executives and the legislators who actually have the authority to change the underlying laws and budgets.
You shift the focus from successfully submitting a single form to successfully delivering the shared outcome.
This is an incredibly daunting task, honestly, but a necessary one. We are moving past the era where we believe digital transformation just meant buying a new software suite. Realising the limitations of the web page, acknowledging that we cannot just slap a responsive website on a broken system and call it fixed, that isn’t the end of digital design. Not at all. It’s a massive evolution. It’s an opportunity to finally focus on what actually matters. The convergence of people, processes and technology around the specific moments in someones life that are genuinely important.
Remember, whatever interface comes next is ultimately temporary. Whether we are interacting with public services through voice-activated assistance, spatial computing or AI-driven chat, that technology will eventually just feel normal to you. It’ll just be everyday life. It will fade right into the background of your daily routine. But what will never fade are fundamental human needs. You will always need immediate care and support when you are sick. You will always need financial security when your life circumstances drastically change. And you will always need the confidence that the institutions built and funded to serve. You can actually orchestrate a response when life gets difficult.
The interface changes. People don’t.
We have to stop trying to build a better candle and start designing what electricity makes possible. Which leaves you with something to really mull over as you look at the apps and sites you use. If we truly stop obsessing over the screen, what does the electricity of services actually look like? Could the future of interaction not be a website, an app, or a form at all? Could it be a completely invisible anticipatory system? A network that knows you just registered the birth of a child or legally incorporated a business and automatically coordinates across a dozen different departments to deliver the tax credits, the health registrations, and the operating permissions you need before you ever even have to ask?
That is electricity.