An AI automation opportunity scanner answers a question every worker and manager now asks: which parts of this role can AI actually do? Pasting a job into a chatbot and asking “can AI do this?” gives mushy answers. This tool reads each task line independently, tags it high, medium, or low for automation potential, and names the category of AI tool that fits — turning a wall of responsibilities into a clear, prioritised map.
How the three levels differ in practice
High automation potential means current AI can handle most of the task with light human review. These are tasks that involve producing text from a template or brief (drafting emails, writing reports), extracting structured data from unstructured sources (forms, invoices, transcripts), classifying inputs into defined categories, scheduling against known rules, or summarising documents. The human’s role shifts to reviewing and approving rather than doing.
Medium automation potential means AI can assist substantially but cannot own the task end-to-end. Research tasks where judgement on source quality matters, customer communications that require relationship context, or data analysis that involves interpreting ambiguous signals fall here. A person still drives the work; AI accelerates it.
Low automation potential means the task requires human presence, accountability, nuanced judgement, or trust that AI cannot reliably supply. Negotiation, counselling, physical inspection, legal sign-off, and managing people through difficult situations are examples. AI might prepare background information, but the task itself stays human.
How it works
You enter a job title for context and paste a task list or description, ideally one task per line. The scanner analyses each line for signals of automatable work: action verbs such as draft, summarise, classify, schedule, transcribe, and extract push a task toward high potential, while verbs implying negotiation, judgement, approval, or physical presence pull it toward low. It then assigns each task a level and suggests a durable tool category — a chat LLM for drafting, a transcription model for meeting notes, a document-extraction model for forms, and so on. The result is a per-task breakdown you can act on immediately. All processing is local to your browser.
What to do with the results
High-tagged tasks are your automation backlog. Before building or buying anything, score each one by frequency (how often it happens per week), time per instance, and complexity. A task that takes 5 minutes but happens 50 times a week is worth automating before one that takes 2 hours but happens monthly. The scanner identifies what can be automated; the frequency and effort analysis tells you what to automate first.
Tips and examples
Split dense paragraphs into discrete tasks before pasting — “manage vendor relationships and process invoices” hides two very different tasks, and separating them lets the scanner tag invoice processing high while keeping relationship management low. Use concrete verbs; “handle reporting” scores worse than “compile the weekly sales report from the CRM”. Treat low-tagged tasks as the human core of the role — the irreducible human contribution — and high-tagged tasks as your automation backlog. After scanning, pair this with a workflow planner to weigh each high-potential task by how often it runs, so you automate the one with the biggest payoff first.