A future we
still get to design.

Boring Future starts with a fairly simple belief: the best technology eventually becomes boring.

Nobody gets excited when a light comes on after flicking a switch. Nobody stops to admire the water infrastructure when they turn on a tap. You do not congratulate your car every morning because thousands of precisely engineered components worked together and the engine started.

It just works.

That is what mature technology looks like. It becomes dependable enough to disappear.

Artificial intelligence is nowhere near that point yet. It is still novel, visible and surrounded by noise. Businesses announce AI strategies. Products are rebranded around it. Every new model release becomes an event. Employees worry about what it means for their jobs while executives wonder how much productivity can be extracted from it.

That will not last forever.

Electricity was once transformative technology. So were telephones, computers and the internet. Eventually they stopped being subjects in themselves and became infrastructure for everything else.

Boring Future is built around the idea that AI and automation should eventually go the same way.

The good version of the future is not one in which people spend their days talking about artificial intelligence. It is one where the report is already there. The customer history has already been gathered. The information needed for the meeting has already been prepared. Two systems that previously required somebody to copy information between them simply talk to each other.

Nobody particularly cares how it happened.

Then the interesting part begins.

What do we do?

That is the future Boring Future is trying to build.

A surprising amount of modern work feels harder than the work itself actually is.

People are employed because of their experience, judgement, relationships, curiosity and ability to solve problems. Yet enormous amounts of their working lives are spent chasing updates, copying information, rebuilding context, filling in forms, maintaining spreadsheets and moving data between systems that should already know how to communicate.

We have somehow turned some of our most capable people into middleware.

A salesperson should spend their time understanding customers, building trust and selling. They should not lose hours filling out information that already exists elsewhere.

A manager should be coaching people, resolving disagreements, setting direction and making decisions. They should not spend half the morning manually assembling the information required to understand what happened yesterday.

An analyst should be interpreting information, finding patterns and challenging assumptions. They should not spend most of their time extracting, cleaning and reformatting the same data before the actual analysis can begin.

People deserve better than that.

So do businesses.

The phrase “Work should feel human” is not simply an argument for making people happier at work. It is an argument about what human beings should actually be spending their time doing.

People should be thinking.

Talking.

Arguing.

Persuading.

Teaching.

Creating.

Making judgement calls.

Changing their minds.

Convincing somebody else to change theirs.

Looking at a process everyone has accepted for a decade and asking the awkward question:

Why do we do it like this?

That is where people become valuable.

And that is where machines should help.

There is something slightly backwards about the way the current AI revolution arrived.

We taught machines to write poetry, create images, make music and imitate many of the things people hold closest to their sense of identity before we fully removed the administrative chores everybody has complained about for decades.

In another timeline, perhaps robotics and automation would have matured first.

The machine would wash the dishes, file the paperwork, reconcile the accounts, move the information and complete the repetitive chores.

Then, much later, the machine would begin writing songs.

Instead, the order has felt almost reversed.

That is part of why AI creates so much anxiety.

People do not merely fear losing tasks. They fear losing the parts of themselves that they recognise in their work: their ideas, their craft, their judgement and their relationships.

Boring Future cannot promise that artificial intelligence will never become capable of doing those things. It probably will become capable of doing many things we currently consider uniquely human.

But capability is not the only question.

The more important question is:

What should we choose to use it for?

Our preference is straightforward.

Give the chores to the machines first.

Keep as much room as possible for the parts of work where having a person involved actually matters.

The useful distinction is not simply between work AI can do and work AI cannot do.

There are really three categories.

Some work is deterministic.

Information needs copying from one place to another. A condition is met and an action should occur. A report needs to run every morning. A form needs routing to the right person. Data needs validating against a clear set of rules.

That is usually a good candidate for automation.

Nobody needs to feel creatively fulfilled by copying the same value into two systems.

Then there is work where the decision belongs to a person but the preparation around that decision is expensive.

Someone needs to read ten documents before making a judgement. A customer history needs summarising. Several options need comparing. Patterns need highlighting. A first draft needs producing.

This is where AI can augment somebody.

The machine does not necessarily make the decision. It gets the person to the interesting part faster.

AI prepares.

People decide.

Then there is work where responsibility, ambiguity, consequence or human relationships mean that the work should remain primarily human.

That boundary should not be based on job titles.

It should be based on whether the whole responsibility can actually be delegated.

An AI may be capable of ranking job applicants. That does not necessarily mean hiring is automatable.

What is the ranking based on?

Could apparently useful correlations introduce discrimination?

Can the system understand genuinely exceptional circumstances?

Who owns the consequences?

Can the organisation explain why the decision was made?

A machine producing an answer is not the same as the entire decision being responsibly automatable.

If a self-driving car cannot reliably identify a red traffic light, we do not describe driving as fully automated and simply choose not to use it. The system has not met the threshold required to perform the job.

Boring Future thinks some business decisions should be viewed similarly.

The question is not:

Can AI technically participate?

It often can.

The question is:

Has the system become capable of carrying the responsibility we are proposing to give it?

Until the answer is yes, a person remains in charge.

Not everything needs AI. Not everything needs automation. The useful skill is knowing which tool fits the work.

Robotic process automation, or RPA, is the rails. Software follows a predefined sequence: open a system, copy an approved value, fill a form, move a file or route an item. It is useful where the path is known and repeatable. The track still needs maintenance: interfaces change, inputs go missing and exceptions need somewhere to go.

AI is the wheels. It can work with language and variation: interpret a request, summarise a document, propose a category or prepare a draft. That flexibility is useful precisely because the input does not always arrive in the same shape. It also means the output needs checking.

The wheels do not get to choose the destination. People define the purpose, the permission boundaries and the point at which a recommendation must become an accountable human decision.

Consider a work order. A rule-based workflow registers it. AI prepares a summary and highlights missing context. A reviewer examines the evidence and approves the next step. An integration then updates the operational record. An unusual or incomplete request goes back to a person.

Use APIs where they provide a dependable connection. Use RPA where an existing interface needs to be operated. Use AI where interpretation earns its place. These are complementary parts of a system, not competing beliefs.

The aim is inwardly automated, outwardly human. Let the system handle the transfer and the preparation, so a person can handle the relationship, read the room and take responsibility.

The best design makes clear where the rails end, where the wheels can move, and who remains in control.

There is another assumption underneath much of the debate around AI: that human beings must somehow justify their continued involvement by remaining more intelligent than machines.

Boring Future rejects that premise.

A rose is not better than a daisy because it is more intelligent.

A tulip is not superior to the grass around it.

A human may possess cognitive capabilities a dog does not, but many people still value, love and admire dogs enormously.

Worth has never been an intelligence leaderboard.

The same is true inside organisations.

The most qualified person in a company does not automatically generate its best ideas.

Someone working on a help desk might notice something that senior leadership has completely missed.

A new starter can be valuable precisely because they do not yet understand why something has “always been done this way”.

They arrive with an outsider's lens.

They learn the process and ask a question everyone else stopped asking years ago.

That perspective can change a business.

Sometimes the best ideas sound stupid at first.

Sometimes the person who sees the problem most clearly is nowhere near the top of the organisational hierarchy.

That is because businesses are not simply collections of intelligence.

They are collections of perspective.

A company is, at its deepest level, a group of people looking at the world together.

Those people bring different backgrounds, experiences, personalities, educations, interests, fears, curiosities, moral boundaries and attitudes towards risk.

One person has spent twenty years in the industry.

Another joined six weeks ago.

One instinctively protects what already works.

Another wants to rip everything apart.

One understands the customer better than anybody else.

Another understands the systems.

Someone is cautious.

Someone is impatient.

Someone has an idea that everybody else thinks is ridiculous.

All of those perspectives collide.

What comes out of that collision is the organisation.

It is a kaleidoscope.

No single colour is inherently more important because it is smarter than another.

Change one piece and the whole picture changes.

This is why leadership matters too.

Apple under Steve Jobs was shaped by a particular personality and set of instincts. Apple under Tim Cook is shaped differently.

We cannot run the alternate timelines and discover what Apple would have become under someone else.

That is precisely the point.

Different people create different organisations.

Even if everyone had access to exactly the same artificial intelligence, they would not necessarily build the same company with it.

And Boring Future believes we should be careful not to flatten that difference.

If increasingly capable general systems begin making more of the decisions, producing more of the ideas and mediating more of the relationships inside businesses, organisations risk gradually outsourcing part of the lens that makes them distinct.

Human participation therefore has value beyond whether a machine could achieve a higher benchmark score.

The question is not only:

Can the machine make this decision?

It is also:

Whose perspective do we want shaping this company?

AI will make some work dramatically faster.

The easy response is to turn every productivity improvement into a headcount calculation.

If ten people can now produce the same output as eight, perhaps the organisation only needs eight.

Sometimes that will be commercially rational. Boring Future is not going to pretend otherwise.

But we think it is the least interesting version of the opportunity.

Imagine a business operating in a £10 billion market and generating £100 million in revenue.

Perhaps AI and automation could allow it to maintain that £100 million with fewer people.

But another question is available.

What could the same people accomplish if 20% of their time suddenly stopped disappearing into administrative work?

Could the company serve customers faster?

Could it follow up opportunities it currently misses?

Could it build the project everyone agrees would be valuable but nobody has time to start?

Could it launch a new service?

Could it improve a process that has been tolerated for years?

Could it move from £100 million to £150 million?

Or £200 million?

There is a floor to cost reduction.

Eventually there is nothing left to cut.

The frontier of what an organisation can attempt is much larger.

That is why Boring Future is interested not only in saved time, but in released capacity.

Saving 2,000 hours is useful.

The more important question is:

What are you going to do with the 2,000 hours?

Most organisations contain an enormous backlog of worthwhile work that simply never gets done.

The customer nobody has time to follow up properly.

The internal process everyone complains about but nobody owns.

The new proposition discussed in three different meetings but never started.

The analysis somebody wishes existed.

The training nobody has time to build.

The improvement project forever scheduled for “next quarter”.

This work does not appear on traditional productivity spreadsheets because it is not happening.

It is invisible.

That means automation can create value beyond replacing an existing activity.

It can unlock activity that previously had no capacity behind it.

This is where Boring Future differs from a traditional automation exercise.

We do not want to stop at:

What can we automate?

We also want to ask:

What becomes possible afterwards?

The ideal project does not merely leave a company doing exactly what it did before, only faster.

It creates room for the organisation to become more capable.

None of this works if employees believe automation is simply a polite way of extracting their knowledge before replacing them.

People are not stupid.

If the implicit message is:

“Explain everything you do so we can work out how few of you we need,”

resistance is rational.

Boring Future believes employees should know when their work is being redesigned.

They should be involved.

In fact, they are usually essential.

Process diagrams rarely capture how work really happens.

The employee knows that the official process says one thing, but everybody actually does something slightly different.

They know that a spreadsheet breaks unless somebody remembers a weird workaround.

They know which customer always needs a different approach.

They know where information disappears.

They know what they hate doing.

They also know what they wish they had more time to do.

So the bargain should be different:

Show us the parts of your job that waste your time.

Then build with the people doing the work rather than around them.

If we can remove enough drudgery, a role may evolve.

Someone may take on more responsibility.

They may finally have time to improve the process rather than merely survive it.

They may discover work they are much better suited to.

That can mean new responsibilities, different job descriptions and sometimes entirely new opportunities.

Automation should be something employees can help shape.

There is also a temptation to interpret every human interaction as inefficiency.

Customer service is an obvious example.

AI systems will become increasingly capable of answering customer questions.

They may be faster than people.

They may know more.

They may be available every second of every day.

That does not necessarily mean every customer interaction should be automated.

Sometimes customers want another person.

Sometimes they want somebody who can listen, understand context and take responsibility.

Sometimes the fact that a person picked up the phone is part of the product.

Trust has value.

Relationships have value.

Perception has value.

The fastest interaction is not automatically the best interaction.

The role of automation may therefore be to make the human conversation better.

Prepare the history.

Surface the relevant information.

Remove the administrative steps afterwards.

Let the person focus on the customer.

That is a much more interesting use of intelligence than simply inserting a bot between two people.

Many businesses already have access to extraordinary technology.

They can buy ChatGPT.

They can buy Claude.

They can buy Copilot.

Their competitors can buy the same things.

Buying a model subscription is not transformation.

The difficult part is the last mile.

How does the intelligence connect to actual systems?

What information can it see?

What should it never see?

What actions can it take?

What rules constrain it?

Where does a person step in?

Who owns the workflow?

How is it monitored?

What happens when it is wrong?

How does somebody change it six months later?

How does it fit into the actual working day rather than becoming another tab people forget to open?

That is where Boring Future lives.

The model is one component.

The value appears when intelligence is connected to systems, context, automation, governance and people in a way that quietly works.

The goal is not to make people think:

“Wow, what an incredible AI system.”

The goal is for them eventually to stop thinking about the AI system at all.

One image captures much of the anxiety around the future of work: skyscrapers full of offices becoming empty.

If artificial intelligence can perform more cognitive work, perhaps eventually there is nobody left inside.

Boring Future does not believe that is the inevitable destination.

The work inside those buildings will change.

Humans once spent enormous amounts of time producing food directly. Most people in developed economies no longer hunt or gather their dinner.

They go to a supermarket.

The ingredients have already been grown, transported, prepared and arranged.

That does not eliminate choice.

Two people can buy the same ingredients and cook completely different meals.

The future of knowledge work may begin to look similar.

Today, people hunt for information.

They gather reports.

They chase colleagues.

They reconcile numbers.

They manually assemble the ingredients required to make a decision.

Increasingly, those ingredients will simply be there.

Prepared.

Organised.

Available.

The interesting question will be what people choose to make from them.

There will still be work.

But the balance can move away from gathering ingredients and towards deciding what meal to cook.

None of this is guaranteed.

Artificial intelligence could absolutely be used primarily to reduce labour costs.

It could mediate more human relationships.

It could centralise decisions inside general-purpose systems.

It could flatten organisational differences.

It could create workplaces where people feel increasingly supervised by systems they do not understand.

Technology does not arrive with a predetermined social outcome attached.

Implementation matters.

Choices matter.

That is why Boring Future is optimistic without assuming everything will simply work itself out.

Businesses need to adopt this technology.

Refusing to change until competitors force the issue is not a serious long-term strategy.

There is truth in the idea that organisations should cannibalise parts of themselves before somebody else does it for them.

But if change is inevitable, there is even more reason to get the direction right while there is still room to choose.

We can make machines better at the chores.

We can use AI to prepare people for decisions rather than quietly removing them from those decisions.

We can increase what businesses are capable of rather than treating every productivity gain as an invitation to shrink.

We can preserve space for disagreement, curiosity, judgement, relationships and individual perspective.

We can build systems that are powerful enough to disappear.

The name is deliberately contradictory.

There is nothing boring about artificial intelligence today.

That is partly the problem.

The technology dominates the conversation.

Boring Future imagines what happens afterwards.

When AI becomes infrastructure.

When automation simply works.

When nobody cares which model prepared the information.

When software handles more of the things people never wanted to spend their lives doing in the first place.

And when people have more room to be different from one another.

Because intelligence was never the entire point.

The point was the salesperson who understands a customer in a way nobody else does.

The new starter who asks why.

The manager who changes their mind after hearing an argument.

The analyst who notices something strange.

The employee whose apparently stupid idea turns out to be brilliant.

The hundreds of individual perspectives that combine into the kaleidoscope we call a company.

Technology should help that kaleidoscope turn.

It should not replace it with one perfect colour.

The best version of the future is therefore not spectacular.

The systems work.

The information arrives.

The repetitive work happens quietly in the background.

The machines do more of the chores.

And somewhere in an office, a group of people are disagreeing about what they should do next.

That is Boring Future.