
Before Eliminating Positions, Understand Which Tasks AI Will Actually Replace
Artificial intelligence is changing the workplace, but one of the biggest mistakes business leaders can make is assuming that every occupation faces the same level of risk.
It does not.
When management asks, “Which jobs will AI replace?”, the question is often too broad.
A better question is:
“Which tasks within each job can AI perform, which tasks should remain under human control, and how should the employee’s role change as a result?”
That distinction is especially important for small and mid-sized businesses, where one employee often performs many different functions.
A person may have a single job title but perform twenty different tasks. AI may eventually automate ten of those tasks, significantly assist with five more, and have very little ability to perform the remaining five.
The job has not necessarily disappeared.
The job has changed.
And that is where thoughtful management begins.
Look at the Back Office Task by Task
Consider a construction, solar, engineering, or service company.
Its administrative employees may spend significant portions of their working day:
- preparing invoices and entering accounting information;
- reconciling routine transactions;
- scheduling maintenance and service calls;
- preparing standard customer communications;
- updating CRM information;
- producing recurring reports;
- processing documents;
- coordinating calendars and meetings;
- creating routine marketing material;
- following up on accounts receivable; and
- transferring information between systems.
Every one of these activities is necessary.
But many share several characteristics.
They are repetitive.
They are information-based.
They follow recognizable rules.
They frequently involve transferring, organizing, analyzing, or communicating information.
Those characteristics make them increasingly suitable for AI-assisted automation.
An AI agent does not necessarily need to replace the employee. It may instead remove hours of repetitive work from that employee’s week.
That distinction is critical.
The Administrative Workday May Become Much Shorter
Imagine an employee who currently spends forty hours each week performing a combination of administrative functions.
Perhaps eight hours are devoted to invoices and accounting entries.
Five hours go toward scheduling.
Five hours are spent preparing reports.
Another five are consumed by routine customer communications and follow-up.
Several more hours involve moving information between email, accounting software, spreadsheets, CRM systems, and project-management systems.
If AI eventually performs much of that work, management may discover that the position no longer requires forty hours of traditional administrative effort.
Perhaps it requires twenty-five.
Eventually it may require fifteen.
That creates an important management decision.
The easiest response is:
“We no longer need this employee.”
But that may also be the least strategic response.
The better question is:
“What should this employee do with the capacity AI has created?”
Automation Should Create Capacity Before It Creates Terminations
Suppose the employee has worked for the company for ten years.
She understands the customers.
She understands how projects move through the company.
She knows which customers require special attention.
She knows the field employees.
She understands the company’s culture and history.
Most importantly, she has demonstrated that she is reliable, adaptable, and committed to the organization.
AI may be able to automate many of her tasks.
But AI has not automatically replaced the value of the person.
The company now has an opportunity to redirect that employee’s time toward activities that create greater value.
She might move into:
project coordination;
service management;
customer relationships;
procurement;
quality assurance;
project controls;
estimating support;
field coordination; or
eventually project management.
This is why I believe companies should think carefully before eliminating good employees simply because portions of their current jobs become automated.
Automate the work that machines can perform. Develop people for the work where people create greater value.
Three Categories of Work
A useful way for management to evaluate every position is to divide its responsibilities into three categories.
- Automate
These are repetitive, rules-based, information-processing activities.
AI and conventional automation should increasingly perform these tasks, with appropriate controls and human review.
Routine data entry, document processing, scheduling, report preparation, and standard communications frequently belong here.
- Augment
These are activities where AI can perform substantial analytical or preparatory work, but people still provide judgment and accountability.
Estimating is a good example.
AI may analyze drawings, compare historical costs, identify quantities, or prepare a preliminary estimate.
But an experienced estimator understands labor productivity, constructability, site conditions, subcontractor risk, customer expectations, and circumstances that may not appear in the data.
The estimator is not necessarily replaced.
The estimator becomes more capable.
The same principle applies to engineering, project management, financial analysis, sales, procurement, and many customer-service functions.
- Keep Human-Led
Some work remains highly dependent on physical execution, accountability, relationships, and situational judgment.
Construction is an obvious example.
AI cannot physically install a solar array.
It cannot replace damaged equipment in the field.
It cannot walk a project site and assume responsibility for safety.
It cannot shake a customer’s hand after a difficult project and rebuild trust.
AI will increasingly assist people performing these activities, but the human role remains fundamental.
Human Review Will Remain Essential
Automation should not mean abandoning controls.
Financial transactions require oversight.
Unusual invoices need investigation.
Accounting classifications may require professional judgment.
Contracts require careful review.
Compliance decisions require accountability.
Customer disputes require context.
Major financial or operational decisions should not simply be delegated to an autonomous system because it is technically capable of generating an answer.
The objective is therefore not automation without people.
It is automation with intelligent governance.
Businesses should determine where human approval is required, who remains accountable, how exceptions are handled, and how AI-generated work is verified.
This will become an important part of corporate governance.
The Real Productivity Opportunity
There is another reason I believe this transition should be viewed positively.
If AI reduces administrative effort substantially, a growing company may be able to increase revenue without increasing overhead at the same rate.
Imagine a company that once required ten administrative employees to support $20 million of revenue.
If AI-enabled systems allow those same ten people—after retraining—to support a significantly larger organization, productivity has increased.
But the greatest opportunity may come when some of those employees move closer to the activities that produce revenue and protect customers.
Instead of adding another administrative position, the company might add:
another service team;
another construction crew;
another project manager;
additional maintenance capability; or
stronger customer support.
AI then becomes more than a cost-reduction tool.
It becomes a capacity-creation tool.
And capacity can create enterprise value.
A Different Conversation With Employees
This transition also requires leadership.
Employees understandably become concerned when they hear that AI can perform portions of their jobs.
Management should not pretend those concerns are unfounded.
Some positions will shrink. Some will disappear. New ones will emerge.
But responsible leadership can make the conversation very different.
Instead of telling an employee:
“AI may replace your job.”
A leader can say:
“AI is going to change part of your job. We want to begin preparing you now for the responsibilities the company will need in the future.”
That is a fundamentally different relationship between technology and the workforce.
It creates an expectation of learning rather than an expectation of displacement.
And it gives good employees an opportunity to participate in the transformation.
Start With a Task Inventory
I recommend that owners and management teams conduct a simple exercise.
Take every administrative and professional position in the company and list the major tasks performed each week.
Do not begin with job titles.
Begin with the actual work.
Then place each task into one of three columns:
AUTOMATE — AI can increasingly perform it.
AUGMENT — AI can make the employee substantially more productive.
HUMAN-LED — physical presence, relationships, accountability, or complex judgment remain essential.
Then ask one additional question:
If AI eliminates 30 percent of this person’s current workload, where can that 30 percent of human capacity create more value for the company?
That may become one of the most important workforce-planning questions of the next decade.
The Special Opportunity for New Owners
This issue deserves particular attention following a management buyout.
New owners understandably want stability after completing an acquisition.
They have debt to service, customers to retain, employees to reassure, and operations to maintain.
But business as usual cannot become the long-term strategy.
The new ownership team has inherited a company built for yesterday’s environment.
Its responsibility is to prepare that company for tomorrow’s environment.
That means understanding how AI will affect every department before competitive pressure forces the issue.
It also means identifying valuable employees today who can become project managers, service leaders, estimators, customer relationship managers, construction managers, and perhaps eventually the next generation of owners.
The Aurora Business Group Perspective
At Aurora Business Group, we believe ownership transition is ultimately about preparing people for greater responsibility.
AI does not change that principle.
It makes it more important.
The companies that navigate this transition successfully will not necessarily be those that replace the greatest number of employees.
They will be the companies that understand their work at the task level, automate intelligently, retain institutional knowledge, develop capable people, and redirect human talent toward activities where it creates greater value.
So before asking:
“Which jobs can AI eliminate?”
Ask:
“Which tasks should AI perform—and what higher-value work can our people perform next?”
Automate the task. Redesign the job. Develop the person.
That is a much stronger AI strategy than simply reducing headcount.
Aurora Business Group
Empowering Employees to Become Owners
