Seth Walsh
Iconoclast
Contributor
- Joined
- Jan 12, 2020
- Posts
- 10,420
- Reputation
- 21,201
The real AI pill is not intelligence. It is attribution.
Most AI threads talk about models.
GPT.
Claude.
Agents.
Coding.
Prompts.
Layoffs.
AGI.
Productivity.
All relevant.
But still surface-level.
The real corporate AI shift is not about who becomes “more productive.”
It is about which parts of the company are allowed to turn productivity into money.
That is the corporate incentive pill.
Inside a corporation, value does not matter until the accounting system can see it.
And the accounting system sees two things very differently:
Profit centres.
And cost centres.
That is the brutal divide.
A profit centre touches revenue.
A cost centre prevents disorder.
A profit centre can say:
A cost centre says:
Both matter.
But corporations do not reward “matter.”
They reward visible contribution to the financial statement.
That is why AI will not hit every job equally.
It will amplify some departments.
It will compress others.
Same model.
Different incentive physics.
The company does not ask if you are intelligent
It asks something colder.
Can I remove this expense without revenue screaming?
That is the question.
Not:
“Is this person smart?”
Not:
“Did this team work hard?”
Not:
“Does this function have dignity?”
The real question is:
What happens to cash flow if this line item shrinks?
That is why cost centres are structurally exposed.
Not because the people are stupid.
Not because the work is fake.
Not because HR, finance, support, legal, ops, admin, compliance, and IT do nothing.
They often hold the company together.
But their success is invisible by design.
When they do their job well, nothing happens.
No lawsuit.
No outage.
No payroll disaster.
No compliance breach.
No angry customer escalation.
No internal chaos.
No executive panic.
That is the problem.
Revenue has receipts.
Prevention has ghosts.
A profit centre gets paid for making something appear.
A cost centre gets paid for making bad things not appear.
AI loves that asymmetry.
Because when AI makes prevention cheaper, the company calls it efficiency.
When AI makes revenue generation stronger, the company calls it growth.
Efficiency gets budget pressure.
Growth gets more budget.
That is the whole game.
AI is not mainly a productivity tool
Inside corporations, AI is a capital allocation test.
It asks:
Who has the authority to turn speed into money?
That is the part employees miss.
Two people can get the same AI tool.
Same subscription.
Same model.
Same laptop.
Same “AI transformation” email from leadership.
One becomes more valuable.
One becomes more replaceable.
Why?
Because one sits on a business loop.
The other sits on an internal artefact loop.
Example.
Person A works in sales.
AI helps him:
If output goes up, revenue can go up.
The company sees it.
CRM sees it.
Pipeline sees it.
Closed-won sees it.
Commission sees it.
Leadership sees it.
Now Person B works in internal operations.
AI helps him:
If output goes up, the company sees fewer delays.
Good.
But what does the CFO see?
Lower need for headcount.
That is the cost centre trap.
The same productivity gain creates different consequences depending on where it lands.
In a profit centre, AI creates expansion.
In a cost centre, AI creates compression.
That is the brutal part.
The cost centre employee adopts AI to become “more useful.”
Then the company uses that usefulness to prove the department can be smaller.
This is why AI feels exciting in revenue teams and existential in support teams.
Same technology.
Different accounting treatment.
Cost centre success is self-liquidating
A cost centre has a hidden curse.
The better it performs, the easier it is to underestimate.
Support team reduces ticket volume?
Great.
Now leadership asks why support has so many people.
Finance automates reconciliation?
Great.
Now leadership asks why finance ops needs the same budget.
HR automates screening, onboarding, policy answers, employee letters?
Great.
Now leadership asks why HR headcount should grow with the company.
Legal automates contract review?
Great.
Now leadership asks which work can move to templates, playbooks, and outside counsel only when necessary.
IT automates provisioning, access requests, password resets, monitoring, internal helpdesk workflows?
Great.
Now leadership asks why internal IT has not become “leaner.”
This is not hatred.
It is incentive logic.
A cost centre is praised when it reduces friction.
Then punished because reduced friction makes it look less necessary.
That is the cost centre paradox:
Your best work removes evidence that you were needed.
A profit centre has the opposite dynamic.
The better it performs, the more evidence it creates.
More revenue.
More users.
More contracts.
More renewals.
More expansion.
More market share.
More pricing power.
It leaves numbers behind.
Cost centres leave silence behind.
And silence never negotiates well.
The real difference is not work
It is attribution.
Corporate power comes from being close to attributed value.
Not effort.
Not hours.
Not intelligence.
Not “being essential.”
Essential but unattributed work gets squeezed constantly.
Attributed value gets defended.
This is why a mediocre revenue generator can be safer than a brilliant internal operator.
Not morally.
Structurally.
The mediocre revenue generator is attached to a number.
The brilliant internal operator is attached to a process.
Numbers win budget fights.
Processes get optimised.
A sales team can say:
“We generated £12 million.”
A product team can say:
“This feature raised retention.”
A pricing team can say:
“This experiment lifted margin.”
A growth team can say:
“This channel lowered acquisition cost.”
A cost centre says:
“We saved time.”
That sounds good.
But time saved belongs to the company.
Not automatically to the employee.
That is the surplus capture question.
When AI makes you 3x faster, who captures the surplus?
If the answer is:
“The company reduces my team.”
You are in a cost-centre dynamic.
If the answer is:
“I close more business, own more scope, control more P&L, or capture more upside.”
You are in a profit-centre dynamic.
That is the entire career pill.
AI does not reward intelligence.
It rewards measurable marginal contribution under authority.
Read that again.
Measurable.
Marginal.
Contribution.
Under authority.
If you are smart but cannot affect a metric, you are decoration.
If you affect a metric but cannot claim it, you are labour.
If you can claim a metric but not control the loop, you are middle management.
If you control the loop, you have leverage.
Profit centres use AI to create more reality
This is what people miss.
Profit centres do not merely use AI to “save time.”
They use AI to run more experiments.
More outreach.
More landing pages.
More product variants.
More pricing tests.
More customer segmentation.
More content.
More demos.
More market research.
More account plans.
More upsell paths.
More conversion loops.
AI increases the number of shots on goal.
And because profit centres are externally validated, the company can see which shots hit.
The market gives feedback.
Customers buy or do not buy.
Users retain or churn.
Accounts expand or contract.
Campaigns convert or fail.
That feedback loop is everything.
Profit centres have external reality checks.
Cost centres have internal satisfaction checks.
External reality compounds.
Internal satisfaction gets budget-reviewed.
This is why AI naturally migrates toward revenue ownership.
Not because sales and product people are smarter.
Because they sit closer to feedback.
Feedback creates learning.
Learning creates better models.
Better models create better decisions.
Better decisions create more money.
More money creates more budget.
More budget creates more AI.
That is the compounding loop.
Cost centres get a different loop.
AI reduces backlog.
Backlog reduction reduces urgency.
Reduced urgency reduces budget.
Reduced budget increases automation pressure.
Automation pressure reduces headcount.
That is the compression loop.
Same company.
Different gravity.
Middle management is the next exposed layer
Before AI, companies needed humans to translate chaos into summaries.
That was real work.
Meetings.
Status updates.
Follow-ups.
Weekly reports.
Roadmaps.
Slide decks.
Alignment notes.
Executive summaries.
Cross-functional handoffs.
“Just checking in” emails.
This created an entire managerial class whose job was not ownership.
It was translation.
Translation between teams.
Translation between workers and executives.
Translation between messy reality and clean slides.
AI makes translation cheap.
Not perfect.
Cheap.
That is enough.
Once summary, coordination, drafting, and reporting become cheap, managers who only move information start looking like overhead.
The future manager has to own one of four things:
Everything else becomes a Slack thread with a salary.
This is why “alignment work” will get brutal.
Companies tolerated coordination bloat when coordination was expensive.
AI makes coordination cheaper.
So now the manager has to prove he is not just a human API between departments.
The old manager asked:
“Can everyone send me their updates by Friday?”
The new manager asks:
“What system makes the update unnecessary?”
That is the jump.
If your job is to collect information, AI is coming.
If your job is to interpret information under uncertainty and make accountable decisions, AI is leverage.
Huge difference.
The new corporate hierarchy
The org chart is becoming fake.
The real hierarchy is not title.
It is loop ownership.
At the top:
In the middle:
At the bottom:
That bottom layer is where the pain will concentrate.
Not instantly.
Not cleanly.
Not in one dramatic robot apocalypse.
It will happen through boring corporate mechanisms.
Hiring freezes.
Attrition.
Shared services.
Vendor consolidation.
Junior roles disappearing.
Teams being “right-sized.”
Internal tools replacing requests.
One person managing what three did before.
Departments being asked to “do more with less.”
Budget not returning after emergency cuts.
That is how corporations actually automate.
They do not announce extinction.
They just stop refilling the chairs.
The AI budget war
Every department wants AI.
But the CFO does not care about enthusiasm.
The CFO cares about which line improves.
A profit centre says:
A cost centre says:
One side speaks in money.
The other speaks in cleanliness.
Cleanliness matters.
But money wins budget fights.
This is why AI ROI feels obvious in some departments and vague in others.
In profit centres, the ROI is additive.
In cost centres, the ROI is often subtractive.
Additive ROI says:
“We made more.”
Subtractive ROI says:
“We needed less.”
Subtractive ROI is politically dangerous because it points back at the team.
The department proves value by proving it can shrink.
That is why cost-centre AI adoption is psychologically strange.
The official story is empowerment.
The financial story is compression.
Both can be true.
The mistake employees make
They ask:
“How do I use AI to do my current job faster?”
Wrong frame.
The better question is:
What business loop am I attaching myself to?
Because doing low-leverage work faster does not make it high-leverage.
It just makes the low-leverage work cheaper.
A person who uses AI to make better internal slides is still downstream of someone else’s decision.
A person who uses AI to own a conversion loop is upstream of money.
Different universe.
Do not become a faster artefact producer.
Become a loop owner.
Not:
“I made a dashboard.”
But:
“I own the metric this dashboard improves.”
Not:
“I automated reports.”
But:
“I removed refund leakage.”
Not:
“I wrote better emails.”
But:
“I increased replies from target accounts.”
Not:
“I improved process.”
But:
“I reduced cycle time from signed contract to cash collected.”
Not:
“I use AI.”
But:
“I use AI to move a number leadership already worships.”
That is how you survive corporate AI.
You move from task identity to metric identity.
Task identity says:
“I do reports.”
Metric identity says:
“I reduce working capital trapped in bad reporting.”
Task identity says:
“I handle support tickets.”
Metric identity says:
“I reduce churn from unresolved customer pain.”
Task identity says:
“I write internal docs.”
Metric identity says:
“I reduce onboarding time for revenue-producing staff.”
Same person.
Different positioning.
One sounds like labour.
One sounds like leverage.
Not all cost centres die
This is where low-IQ AI takes are wrong.
Legal will not disappear.
Finance will not disappear.
HR will not disappear.
Compliance will not disappear.
IT will not disappear.
Operations will not disappear.
The low-agency layers of those functions become software.
The high-agency layers become more important.
There is a massive difference between:
Massive difference between:
Massive difference between:
Massive difference between:
Massive difference between:
AI kills routine mediation.
It does not kill accountable judgment.
The safe person is not the person with a “human” job.
That is cope.
The safe person is the person whose judgment is tied to consequences.
Consequences create authority.
Authority creates budget.
Budget creates survival.
The new elite employee
The new elite corporate employee is not just “good with AI.”
Everyone will be good with AI.
That phrase will become meaningless.
The elite employee has five traits.
1. Metric proximity
He works near numbers leadership already cares about.
Revenue.
Margin.
Retention.
Risk.
Cash flow.
Cycle time.
Conversion.
Customer expansion.
2. System access
He can touch the actual workflow.
Not just comment on it.
Not just make slides about it.
Not just recommend improvements.
He can change the mechanism.
3. Data access
He has access to the proprietary context that makes AI useful.
Customer data.
Operational data.
Product data.
Financial data.
Conversation history.
Internal process data.
Without data, AI is a clever intern with amnesia.
With data, AI becomes institutional leverage.
4. Decision rights
He can act.
This is underrated.
AI amplifies authority more than effort.
A person with decision rights can turn insight into reality.
A person without decision rights can only turn insight into a memo.
Memos are cost-centre artefacts.
Decisions move money.
5. Accountability
He is attached to outcomes.
Not vibes.
Not “supporting the business.”
Not “driving alignment.”
Outcomes.
If the number moves, he gets credit.
If reality breaks, he owns it.
That is the new corporate elite.
Not prompt engineers.
Loop owners.
The final pill
Corporations do not hate cost centres.
They fear unmeasured expenditure.
They do not love profit centres.
They love visible upside.
AI is simply making this distinction more violent.
Before AI, companies tolerated more internal fog.
More manual work.
More coordination drag.
More reporting layers.
More administrative latency.
More “we need another person for this.”
AI attacks fog.
And cost centres contain a lot of fog.
Not because they are useless.
Because their value is often counterfactual, internal, and hard to attribute.
Profit centres have a cleaner story.
Money came in.
The machine likes clean stories.
That is why the AI age will not just divide workers by intelligence.
It will divide them by where their intelligence lands.
Does it land in revenue?
Does it land in risk ownership?
Does it land in systems?
Does it land in data?
Does it land in decision rights?
Or does it land in internal artefacts that help someone else look smart?
That is the line.
Profit centres see AI as a weapon.
Cost centres see AI as a diet plan.
Same model.
Different future.
The question is not whether AI replaces you.
The question is whether your output lands in a revenue line or a cost line.
Because in the age of AI, the corporation will not ask who worked hardest.
It will ask who became cheaper.
Most AI threads talk about models.
GPT.
Claude.
Agents.
Coding.
Prompts.
Layoffs.
AGI.
Productivity.
All relevant.
But still surface-level.
The real corporate AI shift is not about who becomes “more productive.”
It is about which parts of the company are allowed to turn productivity into money.
That is the corporate incentive pill.
Inside a corporation, value does not matter until the accounting system can see it.
And the accounting system sees two things very differently:
Profit centres.
And cost centres.
That is the brutal divide.
A profit centre touches revenue.
A cost centre prevents disorder.
A profit centre can say:
- we created pipeline
- we closed deals
- we increased conversion
- we reduced churn
- we raised prices
- we expanded accounts
- we shipped a product customers paid for
A cost centre says:
- nothing broke
- tickets went down
- reports were delivered
- risk was managed
- employees were supported
- processes stayed clean
- customers complained less
- the machine kept running
Both matter.
But corporations do not reward “matter.”
They reward visible contribution to the financial statement.
That is why AI will not hit every job equally.
It will amplify some departments.
It will compress others.
Same model.
Different incentive physics.
The company does not ask if you are intelligent
It asks something colder.
Can I remove this expense without revenue screaming?
That is the question.
Not:
“Is this person smart?”
Not:
“Did this team work hard?”
Not:
“Does this function have dignity?”
The real question is:
What happens to cash flow if this line item shrinks?
That is why cost centres are structurally exposed.
Not because the people are stupid.
Not because the work is fake.
Not because HR, finance, support, legal, ops, admin, compliance, and IT do nothing.
They often hold the company together.
But their success is invisible by design.
When they do their job well, nothing happens.
No lawsuit.
No outage.
No payroll disaster.
No compliance breach.
No angry customer escalation.
No internal chaos.
No executive panic.
That is the problem.
Revenue has receipts.
Prevention has ghosts.
A profit centre gets paid for making something appear.
A cost centre gets paid for making bad things not appear.
AI loves that asymmetry.
Because when AI makes prevention cheaper, the company calls it efficiency.
When AI makes revenue generation stronger, the company calls it growth.
Efficiency gets budget pressure.
Growth gets more budget.
That is the whole game.
AI is not mainly a productivity tool
Inside corporations, AI is a capital allocation test.
It asks:
Who has the authority to turn speed into money?
That is the part employees miss.
Two people can get the same AI tool.
Same subscription.
Same model.
Same laptop.
Same “AI transformation” email from leadership.
One becomes more valuable.
One becomes more replaceable.
Why?
Because one sits on a business loop.
The other sits on an internal artefact loop.
Example.
Person A works in sales.
AI helps him:
- research accounts
- write outreach
- summarise calls
- prepare proposals
- spot buying signals
- personalise follow-ups
- analyse objections
- forecast pipeline
If output goes up, revenue can go up.
The company sees it.
CRM sees it.
Pipeline sees it.
Closed-won sees it.
Commission sees it.
Leadership sees it.
Now Person B works in internal operations.
AI helps him:
- summarise meetings
- clean spreadsheets
- draft updates
- process tickets
- write internal docs
- reconcile records
- answer repetitive questions
- prepare management reports
If output goes up, the company sees fewer delays.
Good.
But what does the CFO see?
Lower need for headcount.
That is the cost centre trap.
The same productivity gain creates different consequences depending on where it lands.
In a profit centre, AI creates expansion.
In a cost centre, AI creates compression.
That is the brutal part.
The cost centre employee adopts AI to become “more useful.”
Then the company uses that usefulness to prove the department can be smaller.
This is why AI feels exciting in revenue teams and existential in support teams.
Same technology.
Different accounting treatment.
Cost centre success is self-liquidating
A cost centre has a hidden curse.
The better it performs, the easier it is to underestimate.
Support team reduces ticket volume?
Great.
Now leadership asks why support has so many people.
Finance automates reconciliation?
Great.
Now leadership asks why finance ops needs the same budget.
HR automates screening, onboarding, policy answers, employee letters?
Great.
Now leadership asks why HR headcount should grow with the company.
Legal automates contract review?
Great.
Now leadership asks which work can move to templates, playbooks, and outside counsel only when necessary.
IT automates provisioning, access requests, password resets, monitoring, internal helpdesk workflows?
Great.
Now leadership asks why internal IT has not become “leaner.”
This is not hatred.
It is incentive logic.
A cost centre is praised when it reduces friction.
Then punished because reduced friction makes it look less necessary.
That is the cost centre paradox:
Your best work removes evidence that you were needed.
A profit centre has the opposite dynamic.
The better it performs, the more evidence it creates.
More revenue.
More users.
More contracts.
More renewals.
More expansion.
More market share.
More pricing power.
It leaves numbers behind.
Cost centres leave silence behind.
And silence never negotiates well.
The real difference is not work
It is attribution.
Corporate power comes from being close to attributed value.
Not effort.
Not hours.
Not intelligence.
Not “being essential.”
Essential but unattributed work gets squeezed constantly.
Attributed value gets defended.
This is why a mediocre revenue generator can be safer than a brilliant internal operator.
Not morally.
Structurally.
The mediocre revenue generator is attached to a number.
The brilliant internal operator is attached to a process.
Numbers win budget fights.
Processes get optimised.
A sales team can say:
“We generated £12 million.”
A product team can say:
“This feature raised retention.”
A pricing team can say:
“This experiment lifted margin.”
A growth team can say:
“This channel lowered acquisition cost.”
A cost centre says:
“We saved time.”
That sounds good.
But time saved belongs to the company.
Not automatically to the employee.
That is the surplus capture question.
When AI makes you 3x faster, who captures the surplus?
If the answer is:
“The company reduces my team.”
You are in a cost-centre dynamic.
If the answer is:
“I close more business, own more scope, control more P&L, or capture more upside.”
You are in a profit-centre dynamic.
That is the entire career pill.
AI does not reward intelligence.
It rewards measurable marginal contribution under authority.
Read that again.
Measurable.
Marginal.
Contribution.
Under authority.
If you are smart but cannot affect a metric, you are decoration.
If you affect a metric but cannot claim it, you are labour.
If you can claim a metric but not control the loop, you are middle management.
If you control the loop, you have leverage.
Profit centres use AI to create more reality
This is what people miss.
Profit centres do not merely use AI to “save time.”
They use AI to run more experiments.
More outreach.
More landing pages.
More product variants.
More pricing tests.
More customer segmentation.
More content.
More demos.
More market research.
More account plans.
More upsell paths.
More conversion loops.
AI increases the number of shots on goal.
And because profit centres are externally validated, the company can see which shots hit.
The market gives feedback.
Customers buy or do not buy.
Users retain or churn.
Accounts expand or contract.
Campaigns convert or fail.
That feedback loop is everything.
Profit centres have external reality checks.
Cost centres have internal satisfaction checks.
External reality compounds.
Internal satisfaction gets budget-reviewed.
This is why AI naturally migrates toward revenue ownership.
Not because sales and product people are smarter.
Because they sit closer to feedback.
Feedback creates learning.
Learning creates better models.
Better models create better decisions.
Better decisions create more money.
More money creates more budget.
More budget creates more AI.
That is the compounding loop.
Cost centres get a different loop.
AI reduces backlog.
Backlog reduction reduces urgency.
Reduced urgency reduces budget.
Reduced budget increases automation pressure.
Automation pressure reduces headcount.
That is the compression loop.
Same company.
Different gravity.
Middle management is the next exposed layer
Before AI, companies needed humans to translate chaos into summaries.
That was real work.
Meetings.
Status updates.
Follow-ups.
Weekly reports.
Roadmaps.
Slide decks.
Alignment notes.
Executive summaries.
Cross-functional handoffs.
“Just checking in” emails.
This created an entire managerial class whose job was not ownership.
It was translation.
Translation between teams.
Translation between workers and executives.
Translation between messy reality and clean slides.
AI makes translation cheap.
Not perfect.
Cheap.
That is enough.
Once summary, coordination, drafting, and reporting become cheap, managers who only move information start looking like overhead.
The future manager has to own one of four things:
- Revenue — making money appear.
- Risk — being accountable when reality goes wrong.
- Systems — designing workflows that compound.
- Talent density — hiring, developing, and retaining rare people.
Everything else becomes a Slack thread with a salary.
This is why “alignment work” will get brutal.
Companies tolerated coordination bloat when coordination was expensive.
AI makes coordination cheaper.
So now the manager has to prove he is not just a human API between departments.
The old manager asked:
“Can everyone send me their updates by Friday?”
The new manager asks:
“What system makes the update unnecessary?”
That is the jump.
If your job is to collect information, AI is coming.
If your job is to interpret information under uncertainty and make accountable decisions, AI is leverage.
Huge difference.
The new corporate hierarchy
The org chart is becoming fake.
The real hierarchy is not title.
It is loop ownership.
At the top:
- people who own demand
- people who own pricing
- people who own distribution
- people who own customer relationships
- people who own production systems
- people who own proprietary data
- people who own capital allocation
- people who own regulatory risk
In the middle:
- people who turn AI into workflows
- people who connect tools to business metrics
- people who redesign teams around automation
- people who convert messy human processes into scalable systems
At the bottom:
- people who use AI to make prettier artefacts
- people who produce internal documents
- people who attend meetings as output
- people who forward information
- people who maintain process without authority
- people whose work exists only so someone else can decide
That bottom layer is where the pain will concentrate.
Not instantly.
Not cleanly.
Not in one dramatic robot apocalypse.
It will happen through boring corporate mechanisms.
Hiring freezes.
Attrition.
Shared services.
Vendor consolidation.
Junior roles disappearing.
Teams being “right-sized.”
Internal tools replacing requests.
One person managing what three did before.
Departments being asked to “do more with less.”
Budget not returning after emergency cuts.
That is how corporations actually automate.
They do not announce extinction.
They just stop refilling the chairs.
The AI budget war
Every department wants AI.
But the CFO does not care about enthusiasm.
The CFO cares about which line improves.
A profit centre says:
- higher conversion
- bigger pipeline
- faster sales cycles
- lower acquisition cost
- higher retention
- more expansion revenue
- better margin
A cost centre says:
- better employee experience
- faster internal response times
- improved documentation
- more consistent processes
- enhanced knowledge sharing
- reduced manual work
One side speaks in money.
The other speaks in cleanliness.
Cleanliness matters.
But money wins budget fights.
This is why AI ROI feels obvious in some departments and vague in others.
In profit centres, the ROI is additive.
In cost centres, the ROI is often subtractive.
Additive ROI says:
“We made more.”
Subtractive ROI says:
“We needed less.”
Subtractive ROI is politically dangerous because it points back at the team.
The department proves value by proving it can shrink.
That is why cost-centre AI adoption is psychologically strange.
The official story is empowerment.
The financial story is compression.
Both can be true.
The mistake employees make
They ask:
“How do I use AI to do my current job faster?”
Wrong frame.
The better question is:
What business loop am I attaching myself to?
Because doing low-leverage work faster does not make it high-leverage.
It just makes the low-leverage work cheaper.
A person who uses AI to make better internal slides is still downstream of someone else’s decision.
A person who uses AI to own a conversion loop is upstream of money.
Different universe.
Do not become a faster artefact producer.
Become a loop owner.
Not:
“I made a dashboard.”
But:
“I own the metric this dashboard improves.”
Not:
“I automated reports.”
But:
“I removed refund leakage.”
Not:
“I wrote better emails.”
But:
“I increased replies from target accounts.”
Not:
“I improved process.”
But:
“I reduced cycle time from signed contract to cash collected.”
Not:
“I use AI.”
But:
“I use AI to move a number leadership already worships.”
That is how you survive corporate AI.
You move from task identity to metric identity.
Task identity says:
“I do reports.”
Metric identity says:
“I reduce working capital trapped in bad reporting.”
Task identity says:
“I handle support tickets.”
Metric identity says:
“I reduce churn from unresolved customer pain.”
Task identity says:
“I write internal docs.”
Metric identity says:
“I reduce onboarding time for revenue-producing staff.”
Same person.
Different positioning.
One sounds like labour.
One sounds like leverage.
Not all cost centres die
This is where low-IQ AI takes are wrong.
Legal will not disappear.
Finance will not disappear.
HR will not disappear.
Compliance will not disappear.
IT will not disappear.
Operations will not disappear.
The low-agency layers of those functions become software.
The high-agency layers become more important.
There is a massive difference between:
- reviewing template contracts
- and owning legal risk on a strategic deal
Massive difference between:
- processing invoices
- and controlling cash conversion
Massive difference between:
- answering HR policy questions
- and designing compensation systems that attract scarce talent
Massive difference between:
- resetting passwords
- and owning security architecture
Massive difference between:
- checking compliance boxes
- and keeping the company alive in a hostile regulatory environment
AI kills routine mediation.
It does not kill accountable judgment.
The safe person is not the person with a “human” job.
That is cope.
The safe person is the person whose judgment is tied to consequences.
Consequences create authority.
Authority creates budget.
Budget creates survival.
The new elite employee
The new elite corporate employee is not just “good with AI.”
Everyone will be good with AI.
That phrase will become meaningless.
The elite employee has five traits.
1. Metric proximity
He works near numbers leadership already cares about.
Revenue.
Margin.
Retention.
Risk.
Cash flow.
Cycle time.
Conversion.
Customer expansion.
2. System access
He can touch the actual workflow.
Not just comment on it.
Not just make slides about it.
Not just recommend improvements.
He can change the mechanism.
3. Data access
He has access to the proprietary context that makes AI useful.
Customer data.
Operational data.
Product data.
Financial data.
Conversation history.
Internal process data.
Without data, AI is a clever intern with amnesia.
With data, AI becomes institutional leverage.
4. Decision rights
He can act.
This is underrated.
AI amplifies authority more than effort.
A person with decision rights can turn insight into reality.
A person without decision rights can only turn insight into a memo.
Memos are cost-centre artefacts.
Decisions move money.
5. Accountability
He is attached to outcomes.
Not vibes.
Not “supporting the business.”
Not “driving alignment.”
Outcomes.
If the number moves, he gets credit.
If reality breaks, he owns it.
That is the new corporate elite.
Not prompt engineers.
Loop owners.
The final pill
Corporations do not hate cost centres.
They fear unmeasured expenditure.
They do not love profit centres.
They love visible upside.
AI is simply making this distinction more violent.
Before AI, companies tolerated more internal fog.
More manual work.
More coordination drag.
More reporting layers.
More administrative latency.
More “we need another person for this.”
AI attacks fog.
And cost centres contain a lot of fog.
Not because they are useless.
Because their value is often counterfactual, internal, and hard to attribute.
Profit centres have a cleaner story.
Money came in.
The machine likes clean stories.
That is why the AI age will not just divide workers by intelligence.
It will divide them by where their intelligence lands.
Does it land in revenue?
Does it land in risk ownership?
Does it land in systems?
Does it land in data?
Does it land in decision rights?
Or does it land in internal artefacts that help someone else look smart?
That is the line.
Profit centres see AI as a weapon.
Cost centres see AI as a diet plan.
Same model.
Different future.
The question is not whether AI replaces you.
The question is whether your output lands in a revenue line or a cost line.
Because in the age of AI, the corporation will not ask who worked hardest.
It will ask who became cheaper.
McKinsey — State of AI / enterprise adoption and value capture:
https://www.mckinsey.com/capabiliti...w-organizations-are-rewiring-to-capture-value
Gartner — AI value depends on use cases and business outcomes:
https://www.gartner.com/en/articles/ai-value
Fortune / Bloomberg — IBM back-office automation comments:
https://fortune.com/2023/05/01/ibm-ceo-ai-artificial-intelligence-back-office-jobs-pause-hiring/
https://www.mckinsey.com/capabiliti...w-organizations-are-rewiring-to-capture-value
Gartner — AI value depends on use cases and business outcomes:
https://www.gartner.com/en/articles/ai-value
Fortune / Bloomberg — IBM back-office automation comments:
https://fortune.com/2023/05/01/ibm-ceo-ai-artificial-intelligence-back-office-jobs-pause-hiring/
Boardroom image:
G.zengin, Wikimedia Commons, CC BY-SA 3.0
https://commons.wikimedia.org/wiki/File:Boardroom_at_the_Head_Office.JPG
Data centre server racks:
Carl Lender, Wikimedia Commons, CC BY 2.0
https://commons.wikimedia.org/wiki/File:Datacenter_Server_Racks_(22370909788).jpg
Call centre:
Singhira, Wikimedia Commons, CC BY-SA 4.0
https://commons.wikimedia.org/wiki/File:Call_Centre.jpg
Minimalist meeting room:
Breather, Wikimedia Commons, CC0
https://commons.wikimedia.org/wiki/File:Minimalist_meeting_room_(Unsplash).jpg
People in an office:
Minette Lontsie, Wikimedia Commons, CC BY-SA 4.0
https://commons.wikimedia.org/wiki/File:People_in_an_Office.jpg
G.zengin, Wikimedia Commons, CC BY-SA 3.0
https://commons.wikimedia.org/wiki/File:Boardroom_at_the_Head_Office.JPG
Data centre server racks:
Carl Lender, Wikimedia Commons, CC BY 2.0
https://commons.wikimedia.org/wiki/File:Datacenter_Server_Racks_(22370909788).jpg
Call centre:
Singhira, Wikimedia Commons, CC BY-SA 4.0
https://commons.wikimedia.org/wiki/File:Call_Centre.jpg
Minimalist meeting room:
Breather, Wikimedia Commons, CC0
https://commons.wikimedia.org/wiki/File:Minimalist_meeting_room_(Unsplash).jpg
People in an office:
Minette Lontsie, Wikimedia Commons, CC BY-SA 4.0
https://commons.wikimedia.org/wiki/File:People_in_an_Office.jpg
