Answer patient enquiries for a clinic by chat
An n8n and Gemini agent receives patient chat messages, keeps context, uses clinic info from Google Sheets and responds to enquiries.
The author tested Claude computer use in Cowork and Claude Code over two days on six workflows including research, sponsor outreach and product listings.
The author tested Claude computer use in Cowork and Claude Code over two days on six workflows including research, sponsor outreach and product listings.
The author turned on computer use in the Claude desktop app on a Mac and tested six workflows over two days. In the one visible here, she texted Claude through Dispatch from her phone. Claude searched X, Reddit and LinkedIn for posts on her article topic. It put the findings into an Excel file on her Desktop.
Claude returned 44 posts in a two-sheet spreadsheet, with a themes sheet, in about 10 minutes. It captured engagement numbers for LinkedIn but not X or Reddit. The other five workflows are behind a paywall.
The computer must stay on and awake, or Dispatch tasks stop. Computer use needs a Mac and a paid Claude plan.
I run [type of content, e.g. a newsletter or YouTube channel]. Research [topic] by reading the top sources, then build a list of 10 possible sponsors in [niche] with a contact page or email for each, saved in a spreadsheet. Draft a short outreach email for each one, but do not send anything; stop and show me the drafts and spreadsheet for approval first.
An n8n and Gemini agent receives patient chat messages, keeps context, uses clinic info from Google Sheets and responds to enquiries.
An n8n workflow has AI classify web-form jobs by type, urgency and summary, create an Airtable record, reply instantly and flag uncertain cases.
Lindy agents summarized meeting notes, sent action items to teammates, suggested knowledge-base replies on support tickets, and handled follow-up emails and calendar updates.