Last week I demoed our platform to an industry analyst. Near the end, they told me something honest: the research report they are planning will probably not use the words task intelligence, because buyers do not know the category yet. Fair. Their job is labels buyers recognize. Whatever the label, buyers still need to see the work inside a job. Three public accounts raise that question from different directions.
Attributed accounts, not a shared forecast or evidence of Nuvepro delivery.
World Economic Forum, January 2026
Distinguish the purpose of a job from its tasks.
“The question is, what is the purpose of your job?”
Huang used radiology and nursing to argue that automating some tasks can leave more time for patient care.
This is his interpretation of those examples, not a guarantee that every job survives.
Read the World Economic Forum accountThe Adolescence of Technology, January 2026
Task changes can also displace jobs.
Amodei warned that AI could disrupt half of entry-level white-collar jobs over one to five years. His essay treats the speed of adoption and adjustment as uncertain.
A forecast of labor-market risk, not observed displacement or a prediction for one enterprise.
Read Amodei’s essayMorgan Stanley TMT conference, 4 March 2026
Completing a task can create new review work.
Nadella described an AI-generated spreadsheet that handed the work of understanding its output back to him. He discussed using further tools to reason over the result.
The example makes the handoff visible; generating an artifact is not the same as finishing the business work.
Read Microsoft’s transcriptFor a workflow owner, the practical question is more specific than a prediction about jobs: which work should change, who will review it, and how will the result be checked? A generated spreadsheet, a clinical record and an invoice each need their own answer.
Labor economics offers a way to examine activities within jobs.
Task-based economic models, including Acemoglu and Restrepo’s 2019 paper, examine how technology changes the activities involved in production and the demand for labor. They provide a way to ask which work is automated, supported or newly created. They do not make the outcome for a particular job inevitable.
An org chart makes a job visible, but not the work inside it. A financial analyst may prepare a model, check source figures, investigate a discrepancy and explain the result. Those activities need different tools and checks, even when one person owns them.
For our work with enterprises, this makes the activities inside a job worth examining. A title alone does not tell us what an agent would need to do or who should accept its output.
Bring evidence about people, systems and day-to-day work into the same discussion.
A skills profile helps us understand what someone can do. Process records help us trace what systems capture. Interviews, tickets and documents add responsibilities and exceptions that may be missing from either. We use these sources together to understand the selected work.
We call that examination task intelligence. A role carries responsibilities; a workflow connects activities across people and systems. Looking at the tasks inside both helps us specify what an agent should do and where its output must go next. Agents can work across a whole workflow, but the individual decisions and handoffs still need attention.
I will not pretend the name was an instant hit. On sales calls, global systems integrator (GSI) buyers called what we do virtual IT labs, because that is the shelf they knew us from. Analysts filed us next to task mining. For two years the words needed a paragraph of explanation every time I said them. The examples above give me another way to explain why we look inside the job.
Knowing that tasks matter is not the same as knowing your tasks.
A public account can show that a change is possible. It rarely tells you which permissions, records, integrations and exceptions made it work. Those details decide what your own team needs to build.
In our enterprise conversations, a list of job titles is usually easier to obtain than a description of how a ticket, order or invoice moves between teams. That is a useful starting question: where does the work wait, and who handles the cases outside the documented path?
A generic role description can suggest tasks to examine. A manufacturer's automation engineer may have different responsibilities from a software developer in a product company. Compare the actual work before using a role-level score to choose an AI project.
If the task is the unit, the task inventory is the asset. Here is ours.
Our public reference data offers role and workflow patterns to inspect. Labels such as automate, augment and human-only are proposals under stated assumptions. For your organization, review the tasks, available systems and operating rules before choosing an agent project. A catalogue label does not demonstrate that a task has run successfully.
Anthropic’s January 2026 Economic Index analyzes sampled Claude interactions, including patterns of automation and collaboration. That helps describe use on one platform. It does not measure every enterprise’s work or validate the labels in our reference data.
The debate about jobs will run for years. Your move does not depend on who wins it.
You do not need to settle the debate about future employment to examine one current workflow. Choose a problem its owner recognizes, inspect the work behind it, and agree what a useful change would have to demonstrate.
We run this on ourselves. We have mapped our own delivery workflows into their tasks and found tasks no job description owned. After a customer reported slow lab access, we added a user-geography question and a provisioning-region review. We do it live with customers: upload a job description, watch it decompose into classified tasks, argue with the classification, and leave with a proposed preparation plan tied to the tasks that change. That argument, by the way, is the best meeting in the whole process. People defend their work task by task, and that is exactly the conversation that rarely happens when the discussion stays at job titles.
The classification conversation produces questions and a proposed plan. Building the change, preparing its users and checking its performance against agreed cases are further work. We scope that work around the enterprise’s systems and responsibilities.
Tasks, jobs and the limits of applying public examples to your organization.
