![[Non-Engineer's Guide to Claude/Claude Code Series] Cowork Utilization 1: How I Automated After-Hours Information Organization with Cowork](https://devio2024-media.developers.io/image/upload/f_auto,q_auto,w_3840/v1775798367/user-gen-eyecatch/k6ezyklngmwqq06k34s4.webp)
[Non-Engineer's Guide to Claude/Claude Code Series] Cowork Utilization 1: How I Automated After-Hours Information Organization with Cowork
This page has been translated by machine translation. View original
Introduction
Hello! I'm Koga from the Hokkaido Sales Department of the East Japan Sales Headquarters.
I started using Claude one month ago!
This article is an actual experience report on how I, as a sales representative, automated the organization of information that piles up outside business hours using Anthropic's Claude desktop app feature "Cowork."
I want to reduce the time spent on organizing tasks first thing in the morning
As a sales representative in charge of the Hokkaido area, I am responsible for many customers on a daily basis.
As a sales professional, "approaching the customers you should focus on most that day, with certainty." This may seem obvious, but making this a habit and continuing to execute it as part of your morning routine is surprisingly difficult...
Not missing emails that arrived outside business hours.
And reconsidering the priority of follow-up customers based on their content. These two tasks performed right after starting work were part of my daily routine.
The challenge is "unprocessed information that accumulates outside business hours"
Sales days fly by with business negotiations, phone calls, and creating proposal materials. During this time, emails and inquiries from customers continue to arrive. Emails received between the end of the workday and the beginning of the next day often contain information that affects priorities, such as "internal approval has been granted," "need a quote urgently," or "reschedule a meeting."
In other words, the follow-up plan established the previous day may change in priority depending on emails received outside business hours. Only by comparing "CRM status" with "emails received overnight" at the beginning of work can I determine how to proceed that day.
However, this confirmation and reorganization process takes a surprisingly long time, partly because it involves multiple tools. The criteria for "which customer to prioritize" can be inconsistent, and oversights can easily occur. To solve this challenge, I tried the "Cowork" feature of Anthropic's Claude desktop app.
I tried automating "organizing information outside business hours" with Cowork
Cowork is a feature that lets you use Claude as a desktop AI assistant, which can connect to external services (MCP integration) such as CRM and Gmail, and has a scheduling function that automatically executes tasks at specified times. Using this, I built a system that automatically compiles emails received after work hours until before start of work, along with a list of customers to follow up with, and delivers a report timed with the start of business hours.
Overview of the system I set up
| Item | Content |
|---|---|
| Execution timing | Automatically runs on weekdays (Mon-Fri) at the start of business hours |
| Data source ① | CRM: Contacts and deals for the representative/team |
| Data source ② | Gmail: External emails received between end of work and start of work |
| Filter conditions | Long-term no contact (14+ days) or deals due today |
| Priority assessment | Automatically classified into 3 levels: high, medium, low |
| Output format | Markdown file (automatically saved with date) |
| Notification | Notification from Cowork upon completion |
Sample report image
The report automatically generated at the start of business hours consists of the following two sections:
(※ All report contents below are fictional samples)
▼ Follow-up customer list (automatically retrieved from CRM with priorities)
| Priority | Customer name | Company name | Last contact | Recommended action |
|---|---|---|---|---|
| 🔴 High | Company A Representative | (Company name undisclosed) | Yesterday | Prioritize confirmation of XX project progress |
| 🔴 High | Company B Representative | (Company name undisclosed) | 2 days ago | Must reply today due to resend/missed calls |
| 🟡 Medium | Company C Representative | (Company name undisclosed) | 3 days ago | Send service materials |
| 🟢 Low | Company D Representative | (Company name undisclosed) | 5 days ago | Check in when appropriate |
▼ After-hours email summary (external emails received between end of work and start of work)
| Sender | Subject | Received time | Summary |
|---|---|---|---|
| Company A Representative | (Subject undisclosed) | Yesterday 19:00 | Confirmation regarding progress of XX project |
| Company B Representative | [Resend] Schedule adjustment | Yesterday 20:30 | Request to readjust next meeting schedule |
| Company C Representative | (Subject undisclosed) | Today 07:30 | Request for reference materials for implementation |
What particularly impressed me
I was surprised that almost all the setup work could be done in natural language. Just by saying "Target only customers of the team in charge in CRM," Claude identified the representative by searching, and reflected it in the filter conditions. I didn't write a single line of code.
Extension to VSCode (Claude Code)
Since I regularly use Claude Code in VSCode, I also brought the same "team in charge" context to the VSCode side. By describing representative IDs and team information in ~/.claude/CLAUDE.md, I don't have to explain these details repeatedly in conversations with Claude on VSCode.
Effects gained from "simple automation" of work
What I realized through this initiative is that business improvement using AI is "not just for engineers." With Cowork, sales representatives can give instructions to Claude in their own words and achieve automation tailored to their work.
The information organization that I used to postpone at the end of the workday now arrives automatically compiled the next morning.
Just this change has clearly altered the time and concentration available for business negotiations and proposals during the day.
It's a small system, but I feel that it makes a big difference when accumulated daily.
For sales professionals who think "I want to try using AI, but don't know where to start,"
I highly recommend trying Cowork.
The first step is much smaller than you might think!