Handling 83 customer emails an hour without a team
Most service teams waste four hours a day just reading and moving emails. We built a sorter for a small agency in Cambridge that handles 83 messages every hour. This keeps the inbox clean and makes sure the most important tasks stay at the top.
The 4-hour morning slog
In early 2024, a local property maintenance firm near Milton Road was drowning in their inbox. Every single morning, the owner spent about 3.2 hours just deleting junk or moving messages into folders. They were receiving roughly 83 emails every hour during the busy spring season. This meant the office manager had to stay late until 7:15 PM just to reply to basic questions about pricing or availability. It was a total mess because actual urgent repair requests from tenants were getting buried under 47 newsletters and 12 cold sales pitches. The team was stressed and important jobs were being missed.
We sat down with them in March to see if a simple bot could handle the heavy lifting. The goal wasn't to replace the staff but to give them their mornings back. When we looked at the data, 64% of their incoming mail didn't actually need a human to read it immediately. Most of it was just noise or status updates from suppliers that could wait until the afternoon. By the time they finished sorting manually, it was already lunchtime, and they hadn't actually fixed a single customer problem yet. This delay was costing them about £480 a week in wasted admin time.
Sorting by urgency, not by date
The solution we built at TTM Bot wasn't some fancy corporate software. We set up a triage tool that reads the context of the incoming message using simple logic. It looks for specific keywords like 'leak', 'broken lock', or 'emergency' and flags them with a bright red label in Outlook. Everything else gets put into a 'Review Later' folder automatically. This simple change meant the team stopped staring at a screen of 124 unread items every time they sat down. They only saw the three or four things that actually mattered right that second.
By April 2024, the bot was handling about 67% of the initial sorting without a human touching a keyboard. It even started drafting short replies for basic inquiries like 'What are your opening hours on Friday?' or 'Where do I send my invoice?'. (Heads-up: we had to adjust the bot's tone twice because it sounded a bit too formal for a Cambridge trade business at first.) Now, the drafts sit in the outbox waiting for a quick 2-second approval from a human. This keeps the personal touch but removes the 5 minutes it takes to type out the same reply for the hundredth time that month.
I used to wake up to 124 unread messages. Now I wake up to 5 tasks that actually matter.
Getting it right in 21 days
We didn't just turn it on and walk away. The rollout took exactly 21 days from the first meeting to the final hand-off. During the first week, we ran the bot in 'silent mode' where it just watched what the owner did. We compared the bot's choices against 147 historic emails to make sure it wasn't making mistakes. On day 9, we found it accidentally filtered out a legitimate lead from a local council because the subject line looked like a newsletter. We fixed that rule in about 14 minutes and haven't had a false positive since then.
The tech we used is a simple Python script connected to their existing email server. It doesn't require any new logins or expensive monthly subscriptions to massive software platforms. We kept it lean so that if they ever want to change how they work, they aren't locked into a contract. The office manager, Claire, spent about 45 minutes learning how to update the keyword list herself. Now, if they start a new project for 'gutter cleaning' in October, she can add that word to the priority list in three clicks. It is tech that actually does its job without needing a degree to run.
Saving 14 hours every week
The final results were better than we expected for such a small team. After 60 days of running the bot, the agency is saving an average of 14.5 hours of admin work every single week. That is nearly two full workdays recovered. Their response time for 'Urgent' labeled emails dropped from 6 hours down to just 22 minutes on average. This led to a 19% increase in booking confirmations because they were reaching the customers faster than their competitors in the city. They are no longer paying for overtime just to clear an inbox.
Honesty is important here: the bot isn't perfect. About once a week, an email comes in that is so strangely worded that the bot just puts it in the 'General' folder for a human to check. But that is fine. We designed it to be safe rather than aggressive. The main thing is that Sarah and Claire can now spend their time talking to clients and managing their 12 contractors instead of playing digital postman. If you are handling more than 50 emails a day, you are probably wasting time that could be better spent on actual work. We can usually get a triage system like this running for a flat fee.
Our response time for emergencies dropped from 6 hours to 22 minutes without hiring more staff.
