Every automation project and vendor turns up with the same promise: all of the gain, no pain. Install an app, press a button and the organisation digitises itself overnight, without anyone having to change a thing.
Its a lovely idea, at least it would be, if it were true.
There has never been a painless route to real change, and there isn’t one now. Technology might change, but humans are much the same. The friction isn’t a flaw in the process. It is the process. Try to bypass it and you don’t make the pain disappear, you just move it somewhere else, usually onto the people least able to carry it, and usually multiplied.
The upfront work, that slow, awkward business of understanding how things actually get done, is human work. Skip it, and you treat people like parts to be swapped out, rather than the only place your organisation’s hard-won knowledge actually really lives. Lose the people and all that the knowledge goes with them. What you’re left with is a smart new system that falls over the first time it meets a Tuesday. Want to start with nothing? There is a very good reason that startups usually have to raise huge amounts of capital. Inertia is hard, but momentum makes things easier.
We notice this because we spend our time on the part of automation most projects would rather not look at: the people on the receiving end of it, and the services as they are. What you see when you do that depends entirely on where you are standing.
Three lenses on the same project
If your automation programme is stalling, the reason is usually written in the conflicting realities of the people involved.
- The strategy lens. From the boardroom, automation is a lever. A way to lift ROI, run leaner, and grow without raising headcount.
- The technology lens. From IT, it’s plumbing. APIs to connect, models to deploy, and a system that has to stay standing.
- The worker’s lens. From the desk where the work actually happens, it’s a fog. Less certainty, more to hold in your head, and a quiet threat to the one thing you know you’re good at.
None of these views is wrong. The trouble is that most projects pick one and run the whole process from there. The real job is to hold all three at once. That is service design.
The comfort of being good at something
We watch capable people cling to busy work all the time. We’ve seen someone spend the better part of a week hand-collating a spreadsheet that the software could have produced in seconds, if they’d only they would ask and let it.
Why? Because the busy work is a comfort.
The confidence trap
There’s a quiet reassurance in a task you know you can do perfectly, every time. It keeps you in the zone where you feel competent and look productive, which counts for a lot when you suspect your role might be on a list.
Automation asks for the opposite. Learn the new tool. Work in an unfamiliar way. Risk getting it wrong. All where people can see. And it carries a quieter risk on top of that: automation has a way of x-raying the old processes, showing up the ones that looked fine from the outside but were really held together with workarounds, favours and the occasional bit of fudging. None of these things look good.
There’s a hidden cost, too. Automation takes away the brain-dead, repetitive work, leaving only the hard parts behind: the thinking, the judgement, the decisions. Strip out the small, low-stakes jobs that let the mind idle and recover, then hand someone eight unbroken hours of high-stakes thinking instead, and they won’t be twice as productive. They will burn out.
Detecting the signal hidden in the noise
Spend a few decades writing code, testing machine-learning models and reading employee surveys, and you get reasonably good at telling the noise of change from the signal of something actually going wrong.
When someone says “this tool is too hard to use,” a technologist hears a UX bug. A business leader hears a skills or training gap.
But it’s rarely about the interface or the training. The complaint is almost always carrying something heavier underneath.
Sometimes it’s lost agency: “I used to be the Excel wizard. Now I just push a button.”
Sometimes it’s overload: “I don’t need help clicking the button. I need the headspace for the work that comes after it.”
And sometimes it’s fear: “If I get this working perfectly, I’ve automated myself out of a job.”
Listen only to the efficiency numbers and miss all of that, and you’ll drive the organisation off a cliff with the dashboard still showing green.
Giving something back
This is why we build things the way we do at SocialOptic. We start with the people, not the plumbing. Call it coproduction, cocreation, or just plain good design. It is the same process.
We’re not in the business of black boxes that quietly remove roles. We build things that remove the weight sitting on top of people. Free up clinicians from the paperwork so they can get back to patients. Free up teachers so they can get back to teaching. Free up HR teams so they can actually improve the place to work, instead of drowning in the data about it.
Done properly, automation gives something back rather than taking it away. It tells you plainly what’s working and what isn’t, and hands people the evidence to make better calls. We design for trust: you can see what the system did, change it, and stay the expert in the room.
A short manifesto for leaders
If you’re starting down this road, drop that fantasy of the frictionless rollout. Here are three rules for building automation people will actually trust:
Pay for the double-job phase. If you ask your team to document processes, test tools and redesign workflows on top of a full week of their normal manual work, some part of them will quietly sabotage the project just to stay afloat. Clear their plates first. Change isn’t an after-hours hobby.
Watch what’s left on the plate. Human energy isn’t an endless battery. If you automate away 40% of the easy work, don’t backfill the gap with nothing but relentless, high-stakes decisions. Build pauses and recovery into the shape of the day. Connection. The organisation still works off of relationships, however much you automate. They just become higher and higher stakes.
Build a safety net, not a black box. Buy and build tools that put the person in the pilot’s seat rather than the passenger’s. Let them see how the system reaches a decision, and let them overrule it. If people don’t feel safe, the automation is dead on arrival. Human in the loop sounds good, but it ends up with ‘human crushed in the machine’ – instead, empower people.
The pain isn’t optional. You can take it on now, on purpose and in doses you can manage. The alternative is that can wait, and pay it back with interest when the system breaks and your best people are already halfway out the door. You can see the evidence of that out there already.
Photo by Wolfgang Rottmann