Task-level analysis is a method for understanding work by breaking each role into the specific tasks it contains, then assessing each task on its own terms: what skills it needs, what value it creates, and whether AI can do it, assist with it, or barely touch it. It replaces the question "is this job at risk?" with a more useful one: "which parts of this job are changing, and how?"
That distinction is the whole point, so it's worth being concrete about why it matters.
Why job titles are the wrong unit of analysis
Two people with the same job title rarely do the same work. Hire two HR Business Partners into the same band, same description, and within a year one is buried in workforce reporting while the other spends most of their week coaching managers through difficult conversations. Same title. Different jobs.
This was always true, AI just makes it expensive to ignore. When the conversation is "are HR jobs at risk," the answer is either a shrug or a panic, and neither tells you what to do on Monday. When the question is "which tasks in this role are being automated, which are being augmented, and which aren't moving," you get something you can act on: a redeployment plan, a reskilling priority, a redesigned role.
Role-level analysis produces the wrong decisions because it averages away the only information worth having. A role can look "safe" while half its tasks quietly shift to AI, and it can look "at risk" while the tasks that actually define it stay firmly human. The average tells you nothing about either.
What task-level analysis actually involves
The method is straightforward to describe and harder to do well. Four steps:
The five ways a task can change
When we assess a task against AI, it lands in one of five states. This is the framework underneath every GoFIGR assessment:
Most coverage of AI and work collapses this into a binary: automated or safe. Real work doesn't sit at the poles. It sits in the messy middle, and the middle is where the planning happens.
What this looks like in practice
We applied this method to 100 HR roles, broken into 1,818 individual tasks. Even in the most conservative scenario, where an organisation does nothing deliberate about AI, around 85% of those tasks change in some way within three years. Very few get automated outright. The vast majority shift toward "you lead, AI assists" or "AI leads, you guide."
If you'd asked "are HR jobs at risk," the honest answer would have been "mostly no," and you'd have learned nothing. Ask it at the task level and you get the real story: the work is being substantially rearranged, just not deleted, and the organisations that can see which tasks are moving are the ones that can do something about it before it's urgent.
Why this matters more now than it did
Work has always transformed one task at a time. A tool arrives, a task changes, the job title stays the same until it's completely out of step with what the person actually does. What's different now is pace. AI changes the task mix inside a role faster than most organisations refresh a job description, which means the gap between the title and the work is widening for almost everyone at once.
Task-level analysis is how you close that gap deliberately rather than discovering it during a restructure.
We've broken those findings down in detail, by scenario, by HR function, and by the specific skills that survive, in a companion piece: What percentage of HR tasks will AI automate?
Frequently asked questions
What is task-level analysis? It's a method that breaks a role into its component tasks and assesses each one individually, rather than treating the whole job as a single unit. It reveals which specific parts of a role are being automated, augmented, or left unchanged by AI.
How is task-level analysis different from role-level analysis? Role-level analysis asks whether a job is at risk and produces an average that hides the detail. Task-level analysis asks what's happening to each task within the job, which is the information you actually need to redesign work, reskill people, or plan redeployment.
Why does task-level analysis matter for AI workforce planning? Because AI doesn't replace jobs in one move, it changes the tasks inside them. A role can lose half its tasks to automation while its title stays the same. Without task-level visibility, that change is invisible until it forces a crisis.
How do you do task-level analysis? Decompose each role into tasks, classify them, assess each against current and near-term AI capability over a set time horizon, and map the skills each task requires. The output is a per-task breakdown across five impact categories, not a single risk score.
Want to see this for your own role? The free AI Impact Assessment breaks your role down task by task and shows you where each one lands. It takes about three minutes, no signup. Try the free assessment.
If you want to map it across an entire workforce, against your actual AI roadmap rather than industry averages, that's the enterprise assessment.
Helena Turpin spent 20 years in talent and HR innovation where she solved people-related problems using data and technology. She left corporate life to create GoFIGR where she helps mid-sized organizations to develop and retain their people by connecting employee skills and aspirations to internal opportunities like projects, mentorship and learning.
