From Specialists to Versatilists: The Three-Year Roadmap for Organizational Redesign
For over a century, we knew the deal. Specialize deeply, master your craft, climb your functional ladder. The industrial organization rewarded narrow expertise. Finance professionals stayed in finance. Engineers stayed in engineering. The deeper your specialization, the more secure your future.
AI is ending that bargain.
Not because AI is smarter than humans at everything, but because it excels precisely where we've been most rewarded, at repeatable, specialized, knowledge-based tasks.
Keep Reading—and Keep Leading Smarter
Create a free account to finish this piece and join a community of forward-thinking leaders unlocking tools, playbooks, and insights for thriving in the age of AI.
Already a member? Log In
The financial analyst who spent five years mastering complex modeling? AI produces similar analyses in minutes. The marketing specialist who knows every nuance of SEO? AI tools democratize that expertise to anyone with a prompt.
The internet fundamentally transformed how we work with collaboration tools, remote work, real-time information access, but it barely touched how we organize. We bolted new technology onto industrial-era structures. Functional silos remained. Hierarchies stayed intact. Coordination still required endless meetings.
AI won't allow that disconnect to persist.
According to Gartner's recent research presented to CHROs, organizations face a compressed timeline for fundamental redesign: by the end of 2027, leading organizations will have migrated talent reviews away from business units entirely, formally split the manager role, and updated job architectures annually instead of every five years.
Rather than incremental change, what we'll see is a complete reimagining of three core functions of organizational design: what work people do (role design), how decisions get made (decision hierarchies), and how teams work together (coordination).
The question facing HR leaders, COOs, and CEOs is whether they'll proactively lead this redesign as stewards of their people through the most significant technological transformation in over a century, or will they wait until market pressure forces change.
The Three Fundamental Shifts
Shift 1: From Specialists to Versatilists—The Rise of the Conductor
The prediction: By the end of 2028, top-performing organizations will have moved over half their workforce from specialist to versatilist role designs.
Walk into most organizations today and you'll find highly specialized roles designed for task mastery. The Python automation specialist who lives in code. The grocery department shelf stocker who knows one section. The R&D scientist who goes deep in the lab. Each role optimized for narrow, repeatable excellence.
This made sense when coordination across disciplines was expensive. You needed specialists because translation between domains was difficult. Marketing people spoke marketing language. Engineers spoke engineering languages. Collaboration required meetings, documentation, and project managers to bridge gaps.
AI eliminates most of that coordination overhead. It translates natively across domains. It can take a marketing brief and generate technical specifications. It can analyze customer data and suggest product features. The boundaries that made specialization necessary are dissolving.
Enter the versatilist, someone with deep expertise in one or two domains, and broader knowledge of processes and business context adjacent to their expertise. Not a shallow generalist who's mediocre at everything, but someone who can contribute meaningfully across multiple domains and translate between them.
Think T-shaped professionals who've become more comb-shaped: multiple areas of depth with strong connective tissue between them. The Python automation specialist becomes an I&O software engineer who understands infrastructure, operations, and multiple programming paradigms.
The grocery stocker becomes an all-department stocker who understands inventory systems, customer flow, and merchandising. The R&D scientist becomes an R&D product developer who connects lab work to market needs.
The critical role here is what Gartner calls the "conductor", someone who orchestrates diverse capabilities toward outcomes. Like an orchestra conductor who doesn't play every instrument at virtuoso level but understands music deeply, knows what each instrument can do, and knows when to bring in the strings versus quiet the brass.
Organizational conductors:
- Read outcomes and understand what capabilities they require
- Know the humans on their teams strengths, development edges, working styles, and capacity
- Understand what AI can do and when to delegate versus when human judgment matters
- Orchestrate the mix dynamically, shifting who leads based on what the work needs
- Facilitate across differences in discipline, seniority, and personality
- Maintain clarity on outcomes when teams get lost in details
This isn't traditional management. Conductors don't own people's careers, don't do performance reviews, don't control compensation. They temporarily orchestrate capabilities toward an outcome. When that outcome is achieved, the team might dissolve, the conductor might rotate to something else, or team composition might shift.
Current conditions don't incentivize conductors.
In organizations still designed for specialism, adding conductors without changing the underlying structure creates more problems than it solves.
According to Gartner's research, only 42% of employees know how to identify where AI can improve their work, and only 23% say their manager has discussed how their roles will change with AI. When you layer conductor roles onto functional silos, you get conflicts between conductors competing for the same resources, and you risk cannibalizing junior staff who should be developing foundational skills.
The structure itself must change.
The three-year actions
Mid 2026: Incentivize conductor skills
This means more than offering training programs. It requires:
- Rewards and recognition systems that celebrate orchestration, not just individual technical contribution
- Updated competency models and job descriptions that define conductor capabilities
- Compensation premiums for demonstrated conducting ability (Gartner suggests 20-50% premiums for mid to senior level conductors)
- Career paths that don't require becoming a traditional manager
- Public storytelling about successful conducting with specific examples
End 2026: (Re)build expert career paths for foundational skills
Before everyone becomes a versatilist, organizations need people with deep foundational expertise to conduct. This means:
- Identifying which technical disciplines are foundational (the skills conductors will orchestrate)
- Creating clear advancement paths for specialists in these areas
- Ensuring specialists can progress in compensation and influence without becoming conductors
- Building rotation programs where specialists can develop adjacent capabilities
- Making it safe to be a beginner when rotating into new domains
End 2027: Build a process for annual job architecture updates
When entire functions get reimagined repeatedly, organizations need to redeploy human capability quickly. This requires:
- Skills management technology that extracts intelligence from talent and work data
- Quarterly taxonomy reviews to add emerging skills and sunset obsolete ones
- AI-assisted processes that suggest when roles need updating based on actual work patterns
- Governance that allows rapid job description changes without multi-year approval cycles
Shift 2: From Functional Silos to Product-Aligned Structures—Flattening the Hierarchy
The prediction: By 2030, product-aligned structures will surpass functional structures as the most prevalent formal organization design.
The industrial organization was subdivided into manageable spaces: Marketing, Sales, Engineering, Operations. Each function optimized for what human talent could master. Within each function, multiple layers of management aggregated information upward and cascaded direction downward. Seven to ten layers between front line and CEO wasn't unusual.
This worked when information moved slowly and coordination was expensive. Decisions needed to escalate through chains of command. Spans of control couldn't exceed 7-10 people without losing oversight.
AI changes the optimization equation entirely.
Optimizing for human-AI intelligence
When humans alone made decisions, optimal decision rights were structured around functional areas (buying, marketing, selling) and subdivided so talent could make good decisions at the pace of business.
With AI, optimal decision rights are structured around the data variables that explain the most variation in customer outcomes, and they cover wider spaces because humans with AI can make good decisions at faster pace.
Example: A retail merchandise department organized for human intelligence has separate leaders for Men's, Women's, Plus, and Kids, with sub-divisions for Wovens, Knits, Dresses, Denim, and Bottoms under each. That's 15+ separate decision spaces, each sized for human cognitive load.
Organized for human-AI intelligence, the same department reorganizes around age segments that algorithms identify as explaining customer behavior better than traditional categories. Three leaders instead of fifteen. Flatter structure, wider decision spaces.
The result: Fewer layers, faster decisions, outcome-focused rather than function-focused.
Of course, there is a caveat. This requires embedding decision intelligence at the leadership level or it's likely to fail. Does that mean you need a Chief AI Officer? Maybe, but more importantly, all leaders need to experience how much better their decisions are with AI assistance before they'll champion the model throughout the organization.
When executives use AI to analyze market trends, test strategic scenarios, stress-test assumptions, and surface options they haven't considered and then apply human judgment to the final call, they become credible advocates for "human in the loop" decision-making.
The three-year actions
Now: Embed decision intelligence closer to leadership
Forget dashboards and fancy analytics reports, this about changing how leadership teams make decisions:
- Reposition the chief data officer closer to the CEO with direct reporting relationship
- Upskill the data science team's business acumen so they can frame insights for strategic decisions
- Invest in decision intelligence platforms that support scenario modeling, not just reporting
- Start with strategy planning or workforce planning. Use AI assistance but don't announce it widely
- Learn what works, build comfort, then expand to critical decisions
- Make it standard practice for all major decisions, then cascade to next leadership layer
End 2026: Create succession plans with "human in the loop" as core capability
Make AI-augmented decision-making a leadership requirement by:
- Assessing candidates on their ability to effectively use AI to inform decisions while maintaining human judgment
- Looking for evidence of knowing when to trust AI output and when to override it
- Requiring candidates to explain reasoning when diverging from algorithmic recommendations
- Adding explicit criteria: "Demonstrated ability to integrate AI tools into decision-making while maintaining accountability"
- Testing through real scenarios: give candidates complex situations with AI-generated insights and observe how they process information
End 2027: (Re)build organizational muscle at removing leadership blockers
As the organization flattens and the leadership profile changes, some current leaders won't adapt. Organizations need:
- Clear criteria for what constitutes "blocking" the new model (hoarding talent, refusing to use AI tools, optimizing for function over outcome)
- Transparent processes for moving leaders out when they can't evolve
- Executive team alignment that this is necessary, not optional
- Communication that explicitly names the behaviors that no longer serve the organization
- Speed—blockers must be removed within months, not years
Shift 3: From Business Unit Talent Management to Enterprise Communities of Practice
The prediction: By 2029, business units will no longer lead talent reviews or promotion calibrations.
This is perhaps the most radical shift in Gartner's roadmap. For decades, functional leaders owned performance management. The VP of Engineering conducted talent reviews for engineers. The head of Sales decided Sales promotions. Each leader built their bench, protected their high performers, and optimized for their deliverables.
Again, this made sense when functions were kingdoms and career progression meant climbing your functional ladder. But it creates exactly the wrong incentives when adaptability and mobility matter most. Business unit leaders hoard talent and prevent the versatility development that organizations need.
The alternative: separate talent management from work coordination entirely.
Example from Teach For America: TFA reorganized around outcome-aligned structures (fusion teams focused on applicant journey stages), but separated people management into role-specific "chapters."
A software engineer working on the Application team reports to a Software Engineering chapter leader for career development, hiring, firing, and promotion decisions, but takes daily direction from the Application team conductor who's focused on delivering the outcome.
This separation creates different incentives:
- Chapter leaders optimize for long-term capability development across the enterprise
- They have no reason to hoard talent—they don't own outcomes that require specific people
- They can facilitate mobility because moving someone to a different outcome team doesn't threaten their delivery
- Conductors can focus purely on achieving outcomes with the best available capabilities
- People get career development from someone who understands their craft deeply, not just their current project
Skills intelligence infrastructure is a prerequisite.
None of this works if you can't see what skills you have, where they are, and how they're being deployed. Most organizations today know job titles, departments, and reporting relationships.
But ask "who in this company has experience with machine learning?" and you'll get an incomplete list—missing the finance analyst who did ML in graduate school, the operations manager who built predictive models at her last company, and the product manager teaching himself Python at night.
That hidden capability is everywhere, but invisible because organizations track positions, not skills.
The three-year actions
Now: Invest in work and skills intelligence infrastructure
This means:
- Deploying skills platforms (examples: Gloat, Fuel50, Workday Skills Cloud, Eightfold)
- Creating enterprise skills taxonomy with 200-500 granular, meaningful capabilities
- Building real-time skills inventory systems where people update capabilities as they develop them
- Implementing AI-powered skills inference that reads project descriptions, performance feedback, and work samples to identify capabilities even when not self-reported
- Integrating skills data across talent marketplace, learning systems, performance management, compensation, and succession planning
- Making profile completeness part of performance expectations with clear incentives
Mid 2027: Formally split the manager role
This is the pivot point. Organizations must separate:
- Product leadership: Conductors who orchestrate capabilities toward outcomes, handle day-to-day work coordination, make tactical decisions, but do not own people's careers
- People development: Chapter leaders or community of practice leaders who handle long-term workforce strategy, career development, hiring, firing, evaluating, and promoting within their discipline
This split requires:
- Redefining every manager role to be one or the other
- Training both populations on their distinct responsibilities
- Creating new reporting structures where individuals may have different leaders for "what I'm working on" versus "how my career progresses"
- Compensation models that reward both tracks appropriately
- Clear decision rights about who approves what
End 2027: Redesign talent review process from BU-driven to enterprise-driven
Move talent reviews from business units to enterprise communities of practice:
- Reviews focus on capability development, not delivery against functional goals
- Promotion decisions based on skills demonstrated across multiple outcomes, not tenure in one function
- Mobility is encouraged and rewarded, not penalized
- Chapter leaders bring visibility to opportunities across the enterprise
This creates a talent marketplace where:
- Outcome teams put out calls for capabilities
- People see opportunities across the organization
- Matching happens based on skills, interests, and development goals
- Movement between outcomes is fluid
- No one "owns" talent except the enterprise
The Roadmap: What to Do When
The Gartner framework provides a specific sequence that we've broken down in the timeline image below:
This isn't arbitrary sequencing. You can't split the manager role before you have skills infrastructure. You can't migrate talent reviews before you've split the manager role. You can't remove leadership blockers before leaders have experienced the new model themselves.
The timeline is aggressive, three years to fundamentally redesign organizations built over a century. But Gartner's data shows why speed matters: ChatGPT reached 90% adoption in three years. The internet took 23 years. AI adoption is exponentially faster, and organizational design must keep pace.
Why This Matters: Stewardship in the AI Era
The traditional HR conversation about AI focuses on productivity gains, efficiency, and competitive advantage. Those matter. But there's a deeper imperative.
We're entering an age where businesses will need far fewer people to deliver the same output. The question now is whether organizations will simply extract efficiency gains and shed headcount, or embrace a different role as stewards of humanity.
That stewardship requires:
- Cultivating adaptability over narrow expertise, so people can navigate as roles evolve
- Rewarding versatility over specialization, so people build portfolios of capabilities instead of banking on one skill
- Enabling mobility over tenure, so people can move to where value creation is shifting
- Separating worth from job titles, so people's identity and security doesn't collapse when their function gets automated
The three shifts outlined here—versatilists, flattened structures, enterprise talent management—are about creating conditions where humans can continue to contribute meaningfully when machines handle an ever-expanding set of tasks.
The conductor role is fundamentally human: reading context, orchestrating capabilities, facilitating across differences, maintaining outcome clarity. The flattened, product-aligned structure keeps humans close to value creation instead of trapped in coordination overhead. The enterprise talent model ensures people have advocates for their development even as specific outcomes come and go.
The alternative of clinging to industrial org design while deploying AI within it leaves people vulnerable, anxious, and ultimately displaced.
Where to Start
If you're a CHRO, COO, or CEO reading this and thinking "this sounds necessary but overwhelming," here's where to begin:
Month 1: Assessment
- Map your current state against the three shifts. How specialist are your roles? How functional are your structures? Who owns talent decisions?
- Identify which parts of the Gartner roadmap you've already started and where you're furthest behind
- Be honest about leadership readiness—who will champion this and who will block it?
Month 2-3: Build the coalition
- Start decision intelligence practice with executive team on a real strategic decision
- Commission skills intelligence platform evaluation
- Identify 2-3 conductors already operating unofficially in your organization and learn from them
Month 4-6: Pilot
- Launch skills platform with one division
- Test outcome-aligned team structure in one part of the business
- Begin training select leaders in conductor capabilities
Month 7-12: Scale
- Expand skills platform enterprise-wide
- Announce conductor skills as formal career path with compensation premiums
- Put "human in the loop" criteria into next succession planning cycle
- Set date for manager role split
This is a three-year transformation, but it starts with decisions you can make this month.
The internet transformed work but left organizational design untouched. Generative AI will transform both. The organizations that redesign fastest won't just win commercially—they'll fulfill their deepest obligation to the people who make them possible.
