The Day Automation Delivered Data
Yesterday morning, the Claude Code bot running on our Windows device processed buyer data in Google Sheets. The result was clear. Exactly 500 buyers were waiting for a response. These were customers we contacted last month but who have remained silent.
Manual classification would have taken a week. The bot finished in 3 hours. Our team stared at the results for a long moment.
「Okay, so what do we do now?」
The Confusion After Classification
There's a common mistake among automation beginners. When the bot finishes a task, they think the work is done. But reality is different.
Classifying 500 contacts was just the beginning. Here's what we actually had to do.
First Experiment: Re-engagement Email Campaign
Some of the waiting customers might have simply been busy. Among the 500 classified by the bot, we selected 300 who had zero contact over the past 2 months and sent follow-up emails. Different subject lines, simpler content.
Result: 24 people responded within 3 days. The automated classification became the first step in a re-engagement strategy.
Second Experiment: Temperature Segmentation
Instead of treating all 500 the same, we dug deeper. We re-segmented by how long it had been since the last contact.
• 30+ days: 141 people (probably lost)
• 14-30 days: 238 people (high re-engagement potential)
• 7-14 days: 121 people (almost ready to reply)
We applied different message tones and timing to each group. The 7-14 day group got messages on weekend evenings. The 30+ day group got messages tied to new product announcements.
The Most Important Discovery
Something unexpected happened.
60 of the 500 were actually already contracted with other sellers. They told us this by replying to our emails. The bot classified them as waiting to respond, but from their perspective, they had already replied (we just missed it).
This discovery revealed what needed to improve next.
「Classification needs an extra filtering layer」
We made a note on the Windows device. The next bot version will scan email body keywords more precisely and backtrack emails that already expressed response intent but we overlooked.
After Automation
Here's what I learned as a non-developer. Automation returns time to people, but judgment and strategy remain human work.
After the bot classifies, we have more to do, not less. Why didn't this person respond? When should we reach out again? Which message works better? These questions emerge.
Our dream team bots work tirelessly through the night. But looking at their results and asking "what's next" is still our job.
And that's exactly the space automation creates.