What Is a Talent Strategy? How to Build One When AI Is Changing the Roles.

Most companies have a hiring plan. Far fewer have a talent strategy, and the difference shows the first time the plan meets AI. A hiring plan says how many people to add next year and in which teams. It assumes the roles keep their shape.

In 2026 that assumption is the weak point. In a growing number of roles, a real share of the work is moving to software while the plan is still being approved, and the company ends up hiring for jobs that are changing underneath it.

A strategy starts from a different question. Not how many people, but which work people will do, and where each capability will come from.

Talent strategy vs talent management, workforce planning and HR strategy.

Four terms get used for overlapping things, and the overlap is why so many talent strategies turn out to be lists of HR programs. The cleanest way to separate them is by the question each one answers.

Talent strategy makes the choices. Workforce planning puts numbers on them, talent management carries them out, and HR strategy is the wider plan for the function they all sit in.

You will also see "talent management strategy" and "strategic talent management" used for the first row, and in practice they mean the same thing.

The test of the real thing is that it contains decisions someone could disagree with. A document that promises to attract, develop and retain top talent contains none.

Why a talent strategy matters more in 2026.

For most of the last twenty years, the hard part was supply: finding enough of the right people. That problem has not gone away, but a second one now sits on top of it. The roles themselves are changing faster than the plans written for them.

The World Economic Forum's survey of chief people officers at large employers puts reviewing organizational structure and job design at the top of their agenda, chosen by 74%. Their most acute problem is not finding people but matching them to critical skills (World Economic Forum, May 2026).

The scarcity has moved, too. In ManpowerGroup's 2026 survey of more than 39,000 employers in 41 countries, 72% still struggle to fill roles, and for the first time AI skills top the list of what is hardest to find.

The most common response is to retrain the people already on the payroll (ManpowerGroup, February 2026).

CHROs have noticed. One of Gartner's four CHRO priorities for 2027 (October 2026) is to focus on the few skills that set the business apart, as companies thin out management layers and hire fewer juniors while asking for more output. It is the same shift now rewriting HR's own job.

Put together, the message is that you cannot buy your way out. The capabilities you need next are mostly not on the market, and the roles you would hire them into are still being redesigned. That is a strategy problem, not a recruiting one.

The components of a talent strategy.

Strip away the vendor diagrams and a talent strategy framework that works has six parts. The order matters, because each part depends on the one before it.

1. Business outcomes and critical roles. The three to five outcomes the business must deliver in the next two years, and the small number of roles that carry them. A strategy that treats every role as critical has no priorities.
2. Capabilities. What each critical role needs, in skills and in judgment, now and in two years. Judgment is the part most capability maps leave out, and it is the part AI makes more valuable.
3. The sourcing mix. Whether each capability will be built, bought, borrowed or handed to software. Most of the money goes here, and most companies make this choice once, by habit.
4. Development and mobility. How people move into the new work, including the time it takes to stop doing the old work. Internal mobility is the source of talent most companies use least and measure least.
5. Succession and the pipeline. The bench for each critical role, and the early-career intake that will feed it five years from now.
6. Performance, rewards and measures. What gets rated and paid for, and how you will know the strategy is working. If this part contradicts the other five, it wins.

These are what people mean by the pillars of a talent strategy. The seven steps below are how you fill them in.

How to build a talent strategy in seven steps.

Building one is mostly a sequence of decisions. Recruiting, onboarding and development carry the decisions out, but they come after. Here are the decisions, in the order they need to be made.

01Start from the business strategy, not the org chart

Ask the executive team two questions. Which three to five outcomes must the business deliver in the next 24 months? And which roles carry those outcomes, so that if the role sits empty or badly filled, the outcome slips?

The answer to the second question is your list of critical roles, and it is usually shorter than people expect. This is the first step for a reason: everything after it is spent on those roles first. A plan that starts from the org chart ends up spending a little on everything.

02Audit the critical roles for what AI changes

This is the step most plans skip, and it changes every decision after it. In Rewired Leadership I call it the Judgment Audit.

For each critical role, sort the work into three columns: execute, which an agent can take over; synthesize, which the agent and the person do together; and judge, which stays with a person and becomes the core of the role.

Doing the audit honestly means settling which decisions stay with a person and which an agent may take inside clear guardrails. Done well, it shows that the role you are planning for is not the role in the job description. The judge column is where the money should go.

03Map the capability gap

Now compare what the audited roles need with what your people have, in skills and in judgment, for the critical roles only. A company-wide skills inventory is slow to build and stale by the time it is finished.

One across thirty roles can be done in weeks. If the leadership team cannot agree on the size of the gap, an outside assessment of where your people stand today settles it faster than another internal survey.

The gap is usually larger than the leadership team expects and harder to hire for than the recruiting team expects, because the skills it is made of are the ones the market is shortest of.

What a talent strategy can't count on buying: the hardest skills to find in 2026

AI model and application development 20%

AI literacy 19%

Engineering 19%

Sales and marketing 18%

Manufacturing and production 17%

04Choose build, buy, borrow or bot for each capability

Each gap can be closed four ways, known in HR as the four Bs. Make the choice per capability, not per department, and decide bot first, because what an agent takes over changes how much of the other three you need.

Build is the option most employers say they will lean on: retraining is the top response in ManpowerGroup's survey and second only to job redesign in the World Economic Forum's. It only works with a development path designed around the new work, and with time set aside to take it.

05Protect the pipeline

The most expensive mistake available in a 2026 talent plan looks like a saving. The audit shows that much of the junior work is execution, an agent takes it, and the entry-level hiring line gets cut.

That is a loan against the future bench, because the people who would have been your managers in five years were going to come from that intake.

It is already happening. Stanford's Digital Economy Lab, working from ADP payroll records, found that employment of workers aged 22 to 25 in AI-exposed occupations is now 19% below where it would be had it kept pace with less-exposed peers, with no comparable gap for experienced workers.

The shortfall comes mainly from companies hiring fewer young people, not from letting them go (Brynjolfsson, Chandar and Chen, revised August 2026).

The authors call these early indicators rather than proof of cause. For planning, the direction is enough.

The answer is to redesign the entry-level role, not delete it. Give juniors the review of agent output, the escalations and a seat near real decisions, which is the apprenticeship junior work used to provide, in a new form.

The last of the nine questions in the book's CHRO checklist asks exactly this: are we redesigning the early-career pipeline around high-judgment roles that will attract the next generation? It is also the only way to have a bench measured against the roles you will need when the time comes.

06Line up performance, pay and mobility behind it

A strategy that asks for new work while the review form scores the old work will lose, because people follow what they are paid for. Three changes keep the systems pointed the same way as the strategy:

• A review that rewards the judge column. That means a performance system that rates judgment instead of activity.
• An unlearning budget. Funded time for valued people to stop doing work an agent now does and to build the skills for the new work. It tells everyone the company values future capacity over current output.
• Leaders paid for movement. Tie part of a senior leader's bonus to how much their team can decide without them and how many of their people move into higher-judgment roles.

07Measure it and review it every quarter

The strategy needs measures of its own. Time to hire and turnover tell you whether the processes run. They say nothing about whether the strategy is working. Four measures do:

• Critical-role readiness: for each critical role, is one person ready now and another ready within two years?
• Internal fill rate: the share of critical-role openings filled from inside.
• Talent mobility rate: the share of people who move between capability clusters within eighteen months, the book's headline measure of a structure that can change.
• Time to unlearn: how long it takes from an agent taking over part of someone's work to that person working well in a new, higher-judgment role.

Review the four every quarter, next to the business results they serve, and rewrite the strategy when the business strategy changes. That rhythm is the talent strategy roadmap: a short list of decisions, checked every quarter, not a document refreshed every January.

A talent strategy on a page.

A strategy that needs forty slides is a list of programs. A real one fits on a single page. Here is a one-page version: seven questions, with one or two lines of answer each.

1. Outcomes. What must the business deliver in the next 24 months?
2. Critical roles. Which roles carry those outcomes?
3. What AI changes. For each critical role, what moves to execute, synthesize and judge?
4. Gaps. Which capabilities, in skills and judgment, are missing now or will be in two years?
5. Sourcing. For each gap, bot, build, buy or borrow, and why?
6. Pipeline. Who is the bench for each critical role, and how does the entry-level intake feed it?
7. Measures. The four numbers, their targets, and the date of the next review.

Use it as a talent strategy template. Fill it in with the executive team, in the room, before anyone builds a slide. The arguments the page provokes are the strategy.

Talent strategy examples.

Two examples show what these choices look like in practice. The first is illustrative. The second comes from my book.

A customer operations team. Picture a support organization of two hundred people whose plan for next year was to add thirty staff. The audit shows that triage and first drafts of replies are execution, so they go to an AI agent: bot.

Escalations, refunds outside policy and the calls with angry enterprise customers are judgment, and that is where the existing team moves, through a short development path and a quarter of supervised practice: build.

One specialist is contracted to set up the guardrails and quality checks: borrow. The plan still hires juniors, fewer than thirty, but into escalation work instead of triage, so the bench for team leads keeps filling. This example is illustrative, not a client case.

The team that froze. In Rewired Leadership I describe a major digital media company whose senior leaders were committed to empowering their Gen Z employees, and then handed them work without the full context or the guardrails, assuming they would figure it out.

The team stalled. It was not short of initiative. Its members did not know where the boundaries were, and would not risk the company's resources on a path nobody had validated.

When the CEO stepped in, the question changed from why the team had waited to why leadership had failed to provide the guardrails that would make it safe to move. The lesson: hiring for judgment fails without the structure that lets people use it.

Common mistakes.

1. Writing a hiring plan and calling it a strategy. Headcount by team is an output of the strategy, not the strategy.
2. Planning for the roles as they were. If nobody has audited what AI changes, the plan is for last year's jobs.
3. Cutting entry-level hiring because AI does the junior work. It saves a salary now and empties the bench later.
4. Asking for new skills while paying for old output. The pay system wins that argument every time.
5. Treating bot as an IT decision. Which work goes to an agent is a talent decision first.
6. Reviewing the strategy once a year. The roles now change faster than the planning cycle.

Frequently asked questions.

The short version.

A talent strategy is the short set of choices about which work people will do and where each capability comes from: bot, build, buy or borrow. In 2026 it starts by auditing what AI changes in each critical role, keeps the entry-level pipeline alive, and pays for judgment rather than activity.

Everything else is talent management carrying it out. The strategy fits on a page. The hard part is rebuilding the talent processes around the roles you will need, and that is where a strategy succeeds or stalls.