00:00We recently built an AI agent to manage our inbound leads.
00:04The idea was simple, let AI filter out spam inquiries and only pass genuine leads to our team.
00:10Sounds perfect, right?
00:11And for the most part, it worked.
00:13But then something unexpected started happening.
00:16It began marking genuine inquiries as spam.
00:19Real potential customers getting filtered out.
00:22And that's when it hit us.
00:24AI is powerful, but it's not perfect.
00:26Because AI doesn't really understand the contacts the way humans do.
00:31It follows patterns, it follows rules.
00:34And when something slightly different comes in, it can get it completely wrong.
00:39That's why we started thinking about AI very differently.
00:43Not as a replacement, but as a support system.
00:47There is a simple way to look at it.
00:49About 10% should be human setup, defining rules, training system.
00:54Around 80% can be handled by AI.
00:57The repetitive filtering, sorting, processing, etc.
01:01But the final 10% that still need human review.
01:05Especially when it comes to real decisions.
01:08Because one wrong filter can mean one lost customer.
01:12Never mind and be mindful of the future.
01:16No oneillƤ code is pretty famous for you.
01:18You can
Comments