Thank you for sending your enquiry! One of our team members will contact you shortly.
Thank you for sending your booking! One of our team members will contact you shortly.
Duration 7 hours
Course Outline
- Day 1: Delegating Tasks in the Browser
- Mechanisms of Browser Agents
- Processes for reading pages, planning actions, clicking, typing, and verification.
- Distinguishing agent-based browsing from standard chat-based research.
- Current reliability benchmarks and limitations.
- Exercise: Observe an agent perform a simple task and document each step of its operation.
- Comparative Tool Analysis
- Overview of ChatGPT Agent and the Atlas browser.
- Utilizing Claude in Chrome as a browser extension.
- Introduction to Perplexity Comet.
- Considerations regarding availability, cost, and data protection.
- Exercise: Execute an identical research task using two different tools and compare the outcomes.
- Research and Vendor Comparisons
- Aggregating vendor information, prices, and terms across multiple pages.
- Presenting results as a structured table with source links.
- Verifying findings against original source pages.
- Exercise: Generate a vendor comparison based on five specific criteria and validate two data points.
- Forms and Web Applications
- Automating form completion using existing data sets.
- Extracting data from web applications into tabular formats.
- Implementing safeguards to prevent accidental submission, payment, or deletion.
- Exercise: Populate a test form using data from a table and perform a final review before submission.
- Security Protocols
- Mitigating prompt injection via hidden instructions on web pages.
- Managing access to logged-in accounts, passwords, and payment information.
- Utilizing separate browser profiles and approval workflows.
- Exercise: Analyze a pre-configured test page containing hidden instructions and assess the agent's response.
- Establishing Usage Rules and Strategy
- Defining permissible task categories.
- Assigning roles for review and approval.
- Planning for recurring automated workflows.
- Exercise: Draft a one-page rulebook and three practical use cases for your specific team.
- Outcome: Participants will have successfully delegated two tasks to a browser agent, verified the results, and established team-level usage rules.
- Mechanisms of Browser Agents
Requirements
- A laptop equipped with Chrome, an active account for at least one of the featured tools (ChatGPT, Claude, or Perplexity), and a test account containing no sensitive data for exercise purposes.
- No prior programming experience is required.
Target Audience
- Specialist users, assistants, and researchers aiming to delegate recurring web page and application tasks to AI agents.
Testimonials (3)
Comprehensiveness about the backround; how to do a prompt.
Olaf Kannt - Helios Klinikum Krefeld
Course - Generative AI and Prompt Engineering in Healthcare
the tips and recommended prompts that we can take away from this training
Lee Mei Lin - ST Engineering IHQ Pte Ltd
Course - InVideo AI: Creating Engaging Short-Form Videos
use of proper and effective prompt
