For generations, companies have been organized around specialized roles. Marketing generated demand. Sales closed business. Operations delivered. Finance measured the results. Information technology supported the systems everyone else used.
That structure created clarity, but it also created walls.
Artificial intelligence is beginning to weaken those walls. An employee can now research a market, analyze customer data, develop a proposal, create supporting materials, and prototype a new workflow without handing the project from one department to another. The tools do not eliminate the need for expertise, but they allow capable people to contribute well beyond the traditional boundaries of their positions.
That creates both an opportunity and a difficult leadership problem.
Some of yesterday’s strongest employees may be deeply invested in the roles, processes, and organizational boundaries that made them successful. Their expertise remains valuable, but they may resist the very changes the company now needs to make.
Meanwhile, some of the organization’s most future-oriented people may see connections across customers, operations, technology, and strategy that others do not. Yet they may lack the authority, training, or support to act on those insights. They are still being evaluated according to a job description written for a more compartmentalized workplace.
The danger is not simply that companies will adopt AI too slowly. It is that they will fail to recognize where their most important future capabilities already reside.
The archetype beneath the title
A useful way to think about this shift appears in “The Archetype Under the Title”, which discusses a framework proposed by Boris Cherny, creator of Claude Code.
As engineering, product management, design, and data work begin to overlap, Cherny suggests that conventional job functions may become less revealing than five underlying archetypes:
- The Prototyper generates and tests new ideas.
- The Builder turns a promising idea into something dependable.
- The Sweeper simplifies, improves, and removes what no longer serves a purpose.
- The Grower develops an existing product or process until it produces greater value.
- The Maintainer keeps mature systems secure, reliable, and effective.
These archetypes are particularly useful because they do not correspond neatly to departments.
A bookkeeper might be an excellent Sweeper who sees unnecessary complexity in an approval process. A customer-service representative might be a Prototyper who recognizes a recurring customer need and imagines a better service. A salesperson might be a Grower who understands how to refine an offering. An experienced operations manager might be the Maintainer who knows which controls must remain in place as everything else changes.
AI makes it easier for each of these people to move from observation to action.
The customer-service representative no longer needs to wait for a technical team to explore a prototype. The operations manager can investigate data without becoming a programmer. The salesperson can model a new offering, and the bookkeeper can document and test a more efficient workflow.
The employee’s title becomes less important than the combination of judgment, initiative, and working style beneath it.
This does not mean expertise no longer matters
It would be a mistake to conclude that everyone can now do everything.
Deep expertise will remain essential in areas such as finance, law, cybersecurity, engineering, healthcare, and regulatory compliance. AI can help people cross functional boundaries, but it does not remove accountability or turn an enthusiastic beginner into a qualified specialist.
What is changing is the relationship between expertise and execution.
In the past, an employee might identify a problem but have no practical way to explore it outside the boundaries of a department. Today, that person may be able to research the problem, analyze the options, create a preliminary solution, and bring something concrete to the appropriate experts.
The best future employees will therefore be neither narrow specialists who refuse to look outside their functions nor unsupported generalists who underestimate the importance of expertise. They will know when to act, when to collaborate, and when qualified review is essential.
The skills that become more valuable
When AI can perform more of the mechanical work, the remaining human responsibilities become more visible.
The first is judgment. AI can produce many possible answers, but someone still has to recognize which problem is worth solving, which recommendation fits the business, and which risks should not be accepted.
The second is systems thinking. Employees need to understand how a change in one part of the company affects customers, employees, finances, technology, and operations elsewhere.
The third is curiosity. The most useful question may no longer be, “Is this part of my job?” but, “Why does this happen, and who else is affected by it?”
The fourth is communication and influence. Crossing organizational boundaries requires more than generating an intelligent answer. People must listen, explain, resolve disagreements, and earn support from colleagues whose responsibilities and incentives may differ from their own.
The fifth is adaptability. Employees need to move between exploration, implementation, improvement, and stewardship as the work changes.
Finally, they need responsible AI fluency: the ability to use AI productively while recognizing issues involving confidential information, unreliable output, customer and employee data, security, bias, and human accountability.
Technical familiarity can be taught. Judgment, curiosity, trustworthiness, and the ability to bring people together are harder to develop — and increasingly valuable.
How should companies hire now?
Most hiring processes still begin with a familiar title, a list of responsibilities, and a search for someone who has already held substantially the same position.
That approach may fill today’s vacancy while overlooking the people most capable of helping the company change.
Hiring managers should begin asking different questions:
- What kinds of problems does this person naturally notice?
- Do they create ideas, build systems, simplify complexity, grow existing work, or maintain what must remain dependable?
- Can they connect their specialty to the larger business?
- How have they worked outside their formal responsibilities?
- Can they use unfamiliar tools without becoming careless?
- Do they know when to seek expertise or challenge an AI-generated answer?
- Can they bring skeptical colleagues along without dismissing their concerns?
Interviews should place less emphasis on rehearsed descriptions of previous duties and more emphasis on how candidates approach unfamiliar, cross-functional problems.
A useful exercise might give candidates a realistic business situation and ask them to identify the people involved, the information they would need, the risks they see, what they would test first, and how they would evaluate the result. The goal is not to reward whoever produces the slickest AI-generated answer. It is to see how the candidate thinks.
How should companies hire two years from now?
Two years from now, simply claiming proficiency with AI tools will probably distinguish candidates no more than claiming proficiency with email does today.
The more meaningful distinction will be what people can accomplish with those tools — and whether they can do so responsibly with others.
Job descriptions may need to become broader and more outcome-oriented. Companies may hire fewer people solely to execute a fixed collection of tasks and more people to own a problem, workflow, customer outcome, or stage of development.
That does not require immediately abolishing conventional titles. Titles still help define compensation, accountability, professional standards, and career progression. But businesses should avoid allowing those titles to become fences.
A person can remain the director of operations while also being recognized as a Builder. A financial professional can remain accountable for financial work while contributing as a Sweeper across inefficient processes. Archetypes should help leaders understand how people contribute — not become another permanent label limiting what they are allowed to do.
Companies will also need teams with the right balance. An organization full of Prototypers may generate constant excitement but finish very little. Too many Maintainers may protect reliability while missing the need for reinvention. Builders, Growers, and Sweepers are needed to carry ideas through the entire lifecycle.
The hiring question is therefore not only, “Is this a talented candidate?” It is also, “What kind of contribution does our organization need next?”
Start with the people already inside the company
Before rewriting every job description or launching a major recruiting effort, leaders should look carefully at their existing workforce.
Who is already experimenting thoughtfully with AI? Who sees connections between departments? Who has good ideas but no safe way to test them? Who protects essential knowledge and controls? Who simplifies complexity? Who earns trust across organizational boundaries? And who may be resisting change because the company has not explained where they fit in its future?
Resistance should not automatically be interpreted as inability. Experienced employees may see genuine risks that enthusiastic adopters overlook. Their institutional knowledge can be crucial to using AI responsibly.
At the same time, tenure and past performance should not give anyone veto power over necessary change.
Leaders must distinguish between thoughtful skepticism and reflexive protection of familiar territory. They must also give promising employees the training, permission, cross-functional access, and executive support required to contribute beyond their cubicles — physical or virtual.
The real transition is organizational
AI adoption is often approached as a technology decision: choose a platform, purchase licenses, provide training, and wait for productivity to improve.
The larger transition is organizational.
Companies must reconsider how work moves, how decisions are made, how employees collaborate, how talent is evaluated, and which capabilities will matter as traditional roles continue to overlap.
The businesses that handle this well will not discard their experienced employees or chase every new tool. They will combine institutional knowledge with experimentation. They will preserve expertise while reducing unnecessary boundaries. They will identify the Prototypers, Builders, Sweepers, Growers, and Maintainers already in their organizations — and help them work together.
The central workforce question is no longer simply, “Which positions do we need to fill?”
It is: What kinds of people will help this company recognize, build, improve, grow, and sustain what comes next?
Answering that question requires more than a new hiring plan. It requires a company-wide AI transition roadmap — one that connects technology decisions with workflows, workforce readiness, leadership, governance, and business priorities.
The companies that begin that work now will be in a far better position to make the right hires two years from now.