Remember the headlines about companies laying off employees to replace them with AI? Some of those companies are now hiring for the positions they cut after finding that AI could handle only part of the work.
In an article, Forbes describes how misalignments like this occur. For example, when it comes to customer support, AI could answer routine inquiries, while employees were still needed to assess whether a refund was appropriate, recognize when a customer needed more help, and decide when to escalate a problem. Companies had removed people whose responsibilities extended beyond the tasks they were automating. Hiring again meant restoring the human contribution they still needed.
Before making that kind of staffing decision, the business needs to understand what will happen to all the work an employee performs.
Which tasks can AI take on reliably?
Which would benefit from an employee using AI?
Where do experience, relationships, and judgment remain important?
Working through those questions creates room to improve how a team operates without assuming that automating part of someone’s job makes the person unnecessary.
HR can help the business make those distinctions and plan what follows. When tasks change, you need to establish who remains accountable, what employees need to learn, and where their existing skills could be put to better use. Those decisions should shape the proposed changes before employees are asked to adapt to them.
That is where task intelligence in HR becomes useful. Fuel50’s Task Intelligence helps you examine the tasks within roles and consider the appropriate contribution of people and AI. It gives HR a starting point for deciding what to automate, where to support people, and how to prepare employees for the work the organization needs next.
Deciding which tasks to hand to AI also changes what employees are expected to own, how they develop, and where their skills are needed. HR needs enough detail about the proposed changes to help shape those decisions before they become new role descriptions or staffing plans.
Here’s how task intelligence helps with that.
An automation estimate can help identify work to investigate. It leaves important questions unanswered about the responsibilities that remain, the exceptions someone must handle, and how the activities fit together.
Examining tasks individually gives you a way to investigate those dependencies. When routine customer inquiries move to AI, for example, assess the demand for employees who will resolve escalations, check disputed answers, and coordinate with other teams. The work left with people may require a different mix of skills and more time per case.
This is why Fuel50 identifies automation candidates at the task level while keeping the rest of the role visible. The classification helps focus a redesign discussion without turning an entire role into an automation rating.
Technical feasibility tells you only part of what you need to know about a proposed change. You also need to establish who will own the outcome and what experience employees could lose.
A routine case might offer little new learning to an experienced employee while giving a newer colleague practice in interpreting policy and recognizing exceptions. Removing that work creates a development question alongside the efficiency opportunity.
Fuel50’s approach keeps judgment, trust, accountability, relationships, and development visible when assessing the human contribution to a task. This allows HR to examine what should be protected or strengthened as the work changes.
An unchanged job title can conceal a substantial change in expectations. An employee who previously answered routine questions might now be expected to investigate difficult cases, challenge an AI-generated recommendation, or prevent recurring problems.
Those are different responsibilities, with different requirements for proficiency and experience. Task-level analysis helps you specify what employees need to learn and assess whether their existing capabilities could be useful elsewhere.
That connection gives HR a more concrete basis for development and workforce planning. The proposed work determines the capability requirement, which then informs learning, role design, and possible internal moves.
Start with a priority role or team where the business is considering a change. The six steps below apply task intelligence to that decision, using an illustrative customer-support team planning to automate routine inquiries and improve how it resolves more difficult cases.
Work with the business leader to define what better performance would look like. For the support team, the goal might be fewer repeat contacts and faster resolution of difficult problems, with customer satisfaction maintained.
That goal gives you a basis for assessing different arrangements. Automating a first response would be useful only insofar as it helps the team deliver the intended service outcome.
At the leadership level, the CHRO should agree the workforce conditions for the change with business, technology, finance, and risk leaders. Establish who remains accountable, which capabilities the organization must retain, and what support employees will receive.
HR practitioners then turn those conditions into a plan. An HR business partner can coordinate the role review, workforce planners can assess future demand, and L&D can identify the learning and supervised experience required. Technology teams should test what the proposed tools can reliably support, while the business leader owns service delivery.
Include preparation, learning, and supervision time in the proposal. The redesigned team needs resources to reach the expected level of performance.
Ask employees and supervisors to walk through recent cases, including straightforward inquiries and situations that required escalation. For each task, record the intended outcome, the information used, the decisions involved, and any handoff to another person.
A description such as “answer delivery questions” may conceal several activities. In our example, an employee might check the order record, establish whether the information is complete, contact a fulfillment colleague, and decide what to tell the customer. Those activities need to be considered separately before the team agrees what AI could take on.
Distinguish existing AI use from proposed use. Ask where a tool already supports the task, what employees still check, and which problems remain unresolved. Keep assumptions visible until the team has evidence to assess them.
Fuel50’s Task Agency Scale describes four work modes: Automated, Assisted, Collaborative, and Strategically Human. It provides a framework for examining the appropriate balance of human and automation contribution.
Automated: Send a routine order-status response when the information is reliable, the permitted response is defined, and exceptions have a clear route.
Assisted: Help an employee find the relevant policy or prepare a draft response that they check and adapt.
Collaborative: Work through a complex case with AI organizing information and possible responses while an employee questions the evidence and develops the resolution.
Strategically Human: Keep a sensitive complaint conversation human-led where discretion, trust, and the relationship are central to resolving it.
These are possibilities to test in context. A classification does not establish that a particular tool can perform the task reliably in your environment.
For every proposed arrangement, work through whether AI can perform the task, whether it should, whether it already does, and what needs to happen next. That progression reflects Fuel50’s “can it, should it, has it, and what now” approach.
Keep the assessment attached to the work. A high share of automation candidates tells you where to investigate change. It says nothing by itself about an employee’s performance, potential, or suitability for another role. Those questions require separate skills evidence and a conversation with the person.
Give accountability an owner. Specify who checks disputed answers, authorizes exceptions, and can stop or change the process. Give reviewers access to the relevant information and enough time to examine it. “Human oversight” needs to describe work someone is equipped and authorized to perform.
Record the agreed arrangement, the reason for choosing it, the accountable owner, and any unanswered questions. That gives HR and the business a decision they can revisit as the work changes.
Before removing a task, ask what someone learns by performing it and where that learning will happen afterward. The redesigned role should still give employees opportunities to build the judgment and expertise they will need.
For a new support employee, selected routine cases could provide practice in asking clarifying questions, applying policy, and recognizing when an apparently simple request needs investigation. Moving every straightforward case to AI could leave that employee facing complex problems before they have built the necessary foundation.
You do not need to preserve every repetitive step. Identify the capability the experience develops, then design another credible way to build it.
The support team could introduce supervised case reviews. A newer employee examines selected cases, explains which information matters, proposes a response, and discusses their reasoning with an experienced colleague. As their capability develops, they take responsibility for more demanding situations.
Use a mentoring program to support a defined development need, with clear expectations for practice and feedback. Protect time for both the employee and the person supporting them.
The redesign should explain how future employees will acquire the expertise the workflow depends on. Relying indefinitely on the judgment of today’s experienced team leaves that question unanswered.
Once the future responsibilities are clear, specify the skills and proficiency employees need to perform them. Compare those requirements with recent evidence of what people can already do.
In the support team, employees might spend more time investigating repeat complaints or helping improve service processes. Those tasks could require stronger problem-solving, analysis, and cross-functional communication. Training people to use the AI interface would address only one part of the transition.
Build the reskilling and upskilling plan around the actual work. Agree which learning, supervised assignments, and feedback will help close each gap. Define what an employee should be able to demonstrate before taking full responsibility.
Where demand for existing work falls, explore employees’ transferable capabilities and interests alongside opportunities elsewhere. Someone with strong customer knowledge and analytical skills could investigate a service-improvement assignment, provided the work exists and the receiving team can support the move.
Make redeployment part of your internal mobility strategy, with a specific destination, clear requirements, and responsibilities agreed between the current and receiving managers. A short-term project can help both the employee and the receiving team assess the fit.
The transition plan should explain where people could contribute next and how they will prepare. An instruction to “reskill for AI” leaves both questions open.
Return to employees with the proposed design and explain how their input shaped it. Show which tasks will change, which responsibilities remain, and what the organization will do to support the transition.
Be clear about what is decided and what is still uncertain. Where roles may change or disappear, explain how those decisions will be made. Avoid assurances about job security that the organization cannot support.
Make commitments specific enough for employees and managers to use. Identify when learning will happen, who will provide coaching, how internal opportunities will be offered, and where employees can raise concerns about the new workflow.
In the support example, an employee should know whether they will handle more complex cases, receive supervised practice first, or explore a different assignment. Their manager should know how those activities fit alongside service targets.
To build employee trust, make those commitments visible and give people a way to flag when the new arrangement is not working. Asking for buy-in should come with a clear process for addressing concerns and revising the plan.
Test the arrangement before expanding it. Agree a baseline, review dates, and the evidence needed to judge the result. Assess both the service the team delivers and employees’ ability to perform and develop through the new work.
Follow a customer case through to resolution, including AI output, human review, corrections, escalation, and any repeat contact. Compare service quality and total effort with the earlier process.
A quicker initial answer may still lead to more work elsewhere. Check whether employees spend longer correcting responses or whether a small group of specialists now carries more of the difficult cases.
Then examine how any released time is used. For our team, fewer repeat complaints or better resolution of complex cases would provide evidence that the redesigned work supports the agreed goal.
Keep estimated automation potential separate from measured time savings and business results. A task classification does not tell you how much time or money you will save, or whether your staffing needs will change. Measure those outcomes in the redesigned workflow.
Ask managers to review evidence against the revised role requirements. Can employees explain their reasoning, recognize exceptions, and complete the responsibilities they have taken on? Have they had the practice and feedback promised in the transition plan?
Look separately at newer and more experienced employees. Strong team output could conceal a design that depends heavily on a few experts while leaving others without opportunities to develop.
Track whether planned learning, assignments, and internal moves actually happen. Ask employees about workload, clarity of responsibility, access to support, and confidence in raising concerns.
Use those findings to adjust the work allocation, training, or review process before expanding the model. Agree who owns each adjustment and when the team will assess whether it has helped.
Fuel50 Task Intelligence helps HR and workforce teams examine roles through their component tasks. Its task views and classifications provide a starting point for deciding where to investigate automation, strengthen human contribution, or prepare for changing capability needs.
Role-level views show tasks, Task Agency Scale classifications, and role composition. Basic function and business-unit roll-ups help teams examine patterns across a wider area of the organization.
Use those views to focus a discussion with the relevant business leaders. Identify the work that warrants closer examination, validate it with the people who perform it, and agree which proposed changes require testing.
Task Intelligence supports those decisions. It does not implement automation, operate as a process-mining system, or make employment decisions autonomously.
Fuel50’s Skills Intelligence provides the governed skills foundation for understanding capability and role requirements. Its Talent Marketplace supports career paths, learning, mentoring, gigs, and internal opportunities.
HR can use task findings to focus that work on the capabilities and transitions the business needs. A proposed change can inform a role review, a development priority, or a discussion about where existing talent could contribute.
Those follow-on decisions remain human-led. A task classification does not automatically create a learning plan or redeploy an employee. The practical value is a clearer basis for deciding what should change and what support should accompany it.
When you review the next AI proposal, ask for a plan that covers the work people will own, how they will develop, and how the business will assess the result. That gives HR a meaningful role in shaping the redesign before its consequences become an employee transition problem.
Explore Task Intelligence with Fuel50 and discuss how to apply this approach to a priority role in your organization.