← back
Beyond APIs: How AI Web Agents Are Automating the "Long Tail" of Knowledge Work
Takeaway
Browser-side AI web agents like Retriever automate the long tail of knowledge work where APIs don't exist — multi-tab scraping + dynamic third-party tool calls.
Summary
- Retriever (retriever.com) is a Chrome extension AI web agent that automates browser workflows — copy-paste between sites, scrape to Google Sheets, fill forms, follow pages.
- Sub-cent cost per page extraction; handles multi-tab parallel processing — give it a Google Sheets column of URLs and it opens/processes each independently.
- Deep-research feature explores multiple pages per company (homepage → pricing page) and self-corrects wrong URLs (typed company.com but agent navigates to Yahoo Finance for P/E ratio).
- Dynamic function calling — point at any third-party API spec (e.g. WhatsApp integration) and the agent invokes it per row in a sheet (sent test messages successfully on stage).
- Built-in graph-bot mini-agent generates data-analysis charts on extracted data.
web-agentsbrowser-automationretriever
Original description
AI Web Agents, such as rtrvr.ai, that can autonomously take actions on the web, scrape data, create graphs, and call API's with just natural language represent a fundamental reimagination of what we can do with the browser. https://www.rtrvr.ai/blog/introducing-rtrvr-ai