Your Business Has Too Many Humans in the Loop
and no, I’m not suggesting we get rid of them. There, I said it!
Your business probably has too many humans in the loop but before someone screenshots that sentence and announces that Steve AGI wants to replace everyone with robots…
I don’t. Quite the opposite, in fact!
I think humans are far too valuable to spend their lives doing work that doesn’t actually require a human.
Pumping Gas, punching data, collecting leeches (yes it was once a thing) or simply mind numbingly boring admin; yet that’s exactly how most businesses choose to operate, apart from your local leech collector service!
I call it the Human Middleware Problem
A customer fills out a form.
Someone reads it.
Someone copies the information into another system.
Someone emails another person.
That person checks a spreadsheet.
Someone approves something.
Someone else updates the CRM.
Then somebody remembers to follow up three days later.
Seven human interactions.
One fairly ordinary process, and somewhere there’s a manager proudly saying:
"We have strong controls."
Mate! You don’t have controls.
You have expensive middleware wearing name badges, because we’ve confused Human Involvement with Human Control
This is an important distinction.
Humans should absolutely remain in control of important decisions, but humans don’t need to physically touch every step of every process to maintain that control.
Think about a payment.
Does a human need to manually process every normal payment? No.
Once upon a time they did, now? Nope!
Should a human investigate an unusual $47,000 deposit to Libia transaction?
Probably a bloody good idea. That’s the model.
Systems handle the normal. Humans handle the exceptional.
Yet most businesses have built the exact opposite, where everything goes through people.
Normal.
Abnormal.
Important.
Trivial.
So genuinely important decisions compete for attention with rubbish.
That’s Not Governance; It’s congestion and regulated industries are particularly guilty of it.
Healthcare.
NDIS.
Finance.
Government.
Agriculture.
Insurance.
Someone inevitably says:
"We can’t automate that. We’re too regulated."
Sometimes they’re right, often they’re not.
Regulation usually tells you what controls, evidence and accountability must exist.
It doesn’t necessarily say Karen has to copy something from Outlook into Excel every Thursday afternoon. We’ve somehow turned compliance into a justification for bad process design.
Enter AI, which is where AI gets genuinely useful.
Not:
"Let’s replace Karen."
But:
"Why is Karen doing that in the first place?"
That’s a much better question.
AI and automation can increasingly handle the operational layer around people.
Capture the information.
Classify it.
Check it.
Route it.
Summarise it.
Recommend an action.
Execute approved actions.
Escalate exceptions.
Record what happened.
Suddenly the human isn’t removed.
The human is elevated. Instead of processing everything, they’re deciding what matters.
The Exception Economy
I think this is where business is heading, unless they’re on a road to nowhere…
People won’t manage processes. They’ll manage exceptions.
The ordinary stuff will increasingly disappear into systems.
Customer enquiry comes in?
Handled.
Invoice due?
Followed up.
Appointment booked?
Confirmed.
Standard application?
Processed.
Stock threshold reached?
Action triggered.
Something unusual happens? Human.
Something high-risk happens? Human.
Something requires empathy? Human.
Something requires judgement? Human.
Something genuinely matters? You guessed it…drum roll please…Human.
That isn’t replacing people.
That’s finally using people properly and this changes management too
Here’s where it gets interesting.
Most managers spend enormous amounts of time checking whether things happened.
Did we follow up?
Did someone respond?
Was the report completed?
Did that customer get called?
Has the invoice gone?
That’s not leadership; That’s human monitoring software. A properly designed system should already know., wich means managers can stop policing activity and start improving outcomes.
Imagine that, managers actually managing, revolutionary stuff.
What about me? (by ‘me’ I mean you!)
This is one of the questions we now ask when designing KAOS AI Decision Systems.
Not:
"Where can we use AI?"
But:
"Where must a human remain in control?"
That’s deliberately backwards, because once you’ve defined that boundary, everything else becomes much clearer.
That’s part of what we’re now calling the KAOS 6:
WORKFLOW What exactly are we improving?
OWNER Who is accountable?
SIGNAL What causes something to happen?
DATA What information can we trust?
CONTROL Where must a human remain involved?
OUTCOME What measurable result should change?
Notice what’s missing?
"Which AI model should we buy?"
That’s intentional…Technology comes later.
I’m becoming increasingly convinced that most AI projects are too bloody big.
"AI transformation."
"Enterprise AI strategy."
"Intelligent organisation."
Lovely PowerPoint.
See you in 18 months.
I’d rather take one annoying workflow and make it demonstrably better in 30 days.
One workflow.
One owner.
One outcome.
Can we:
Cut response time from four hours to four minutes?
Reduce administration by ten hours a week?
Recover more abandoned quotes?
Reduce missed appointments?
Spot problems earlier?
Great! Prove it, then expand!
The Real Future of Work is clear ; we’re spending far too much time asking:
"Which jobs will AI replace?"
I think the more useful question is:
"Which parts of every job should disappear?" – Because almost every role contains work that deserves to survive…and work that absolutely doesn’t.
The nurse should nurse.
The salesperson should sell.
The advisor should advise.
The manager should lead.
The farmer should farm.
The founder should build.
And the systems? Well, they should handle as much of the repetitive operational crap surrounding those people as safely possible.
My Prediction is that the best businesses of the next decade won’t be completely automated and they won’t be completely human.
They’ll be deliberately human. Human by choice, by design; not by default!
They’ll know exactly where people create disproportionate value.
Trust.
Judgement.
Creativity.
Empathy.
Relationships.
Leadership.
And they’ll design everything else around protecting those moments.
That’s the future I find considerably more interesting than another AI chatbot, but hey, that’s just me
Thought of the Week
You don’t need to remove humans from business; you need to:
Take away the chasing. The copying. The remembering. The checking. The unnecessary approvals. The administrative sludge. Then see what your people can actually do.
AI IS OK, not because businesses need fewer humans, but because businesses need humans doing more human work.
Less processing. More judgement.
Less admin. More impact.
Less chaos. More control.
Steve AGI
Founder | C51 Consulting Architecting AI Decision Systems
#AIisOK • #CONTROLtheCHAOSwithKAOS • #AGIsGotThis