Alex RaineyAlex is an experienced CTO and founder who largely focuses on all the technical areas of My AskAI, from AI Engineering, Technical Product Management and overall Platform Development.
Created time
Aug 25, 2024 08:31 PM
Author
Publish date
Aug 25, 2024
Slug
how-we-scaled-ai-customer-support-to-handle-71617-conversations
Featured
Featured
Type
Article
Ready to Publish
Ready to Publish
In the last 30 days, we processed 71,617 customer support conversations.
Of these, over 55,000 were resolved by our AI agent (77%). A year ago, it was a few hundred/month. And we could only answer ~50% of customer queries (on a good day).
When you start building a SaaS product, some things just aren't a priority:
- Scalability
- Performance
- Graceful fallbacks
- Automated tests
- Uptime alerts
Then suddenly one day, you realize that large companies rely on you for a core part of their business. We have some customers with a few hundred conversations per month, but we have some with 20k+/mo. Any disruption to our service causes serious issues for their human agents.
So I spent the last week purely focusing on stability tasks. They're not sexy. They're not that fun to build. Most people will never see orfeelthem. But that's also the point.
We’ve been able to successfully reduce our reliance on 6 providers down to 2. This means we can keep running even when 70% of our core third parties go down.
- OpenAI down, no problem.
- AWS or Azure instance fallen over, no worries.
This is the right thing to do for our customers, but it also gives us such peace of mind. As a small team, having a resilient tech stack is vital at our stage.
Written by
Alex Rainey
Alex is an experienced CTO and founder who largely focuses on all the technical areas of My AskAI, from AI Engineering, Technical Product Management and overall Platform Development.
