Browser Automation Task Spec Prompt
Converts a manual click-path into a reliable automation spec.
/ quick answer
Use before building any browser agent so the model isn't guessing the flow. Converts a manual click-path into a reliable automation spec.
You are a browser automation engineer. I will describe a manual task performed in a web portal. Produce:
1. STEP MAP — each step with the element to interact with and the expected page state after it.
2. DETERMINISTIC vs MODEL — mark which steps should be hardcoded selectors and which genuinely need a model, with reasoning.
3. FAILURE POINTS — what breaks (modals, session expiry, pagination, rate limits) and the recovery for each.
4. VERIFICATION — the assertion that proves the task actually completed.
5. SECRETS — credentials needed and how they should be injected.
6. ESCALATION — when to stop and hand a screenshot to a human.
Task: {{TASK}}
Portal: {{PORTAL}}
Frequency: {{FREQUENCY}}1. STEP MAP — 1) Go to /login, expect email field. 2) Fill credentials, expect dashboard heading...
What does the Browser Automation Task Spec Prompt prompt do?
Use before building any browser agent so the model isn't guessing the flow.
Which AI models work with this prompt?
It is model-agnostic: it works with any capable general model. Replace the bracketed variables with your own context before running it.
What output should I expect?
1. STEP MAP — 1) Go to /login, expect email field. 2) Fill credentials, expect dashboard heading...
/ continue exploring
Related concepts
The vocabulary this page depends on.
- →Browser Agent
A browser agent drives a real web browser — navigating, clicking, typing and reading pages — to complete tasks on sites that have no API.
- →Computer-Use Agent
An AI agent that controls a desktop or browser via screenshots and clicks.
- →No-Code Automation
Building business workflows visually without writing code.
- →Research Automation
Research automation turns a question into a sourced, structured answer using search, retrieval, extraction and synthesis agents.
Related workflows
Turn this into a repeatable process.
- →Automate a Portal That Has No API
Use a browser agent where integration is impossible — without daily breakage.
- →Build a Research Automation Pipeline
Question in, sourced structured brief out — on a schedule.
Related tool stacks
The tools that run it in production.
- →Browser Automation Stack
Run browser agents on a schedule with credentials, retries and screenshots.
- →Research Automation Stack
Search, fetch, extract and synthesise sourced briefs on a schedule.
- →Crypto Automation Stack
Automation platform, data APIs, alerting and optional wallet execution — the operational layer for monitoring and recurring actions.
Related prompts
Reusable prompts for this job.
- →No-Code Automation Spec Writer
Turn a vague 'I want to automate X' into a buildable scenario spec for Make / n8n / Zapier.
- →Sourced Research Brief Prompt
Produces a structured brief where every claim carries a citation.
- →Document Extraction Schema Prompt
Designs the schema and validation rules before you build extraction.
- →Coding Agent Task Spec Prompt
Writes the task brief that keeps a coding agent inside scope.
Related use cases
How people apply it, and what came out.
- →Finance Team Automates 14 Supplier Portals with Browser Agents
Weekly invoice retrieval from portals with no API or export.
Comparisons & alternatives
Pick between the options.
- →Browser Agent vs API Automation
APIs win whenever they exist; browser agents exist for the systems that never gave you one.
- →Best AI Workflow Automation Tools: n8n vs Zapier vs Make
The three tools most operators consider for AI workflow automation — compared on pricing, AI integration and technical flexibility.