Our AI outbound efforts were a black box. Here’s the real cost.

There’s a specific kind of pressure that comes when your commercial team hits a wall.

Despite being swamped with managing current clients, pitching to new leads, creating proposals, and closing deals, the team at Tech in Asia knew we needed a contingency plan to secure revenue targets.

We couldn’t just rely on existing conversations and wait for inbound interest to pick up. We had to do something. However, building a proper cold outbound engine from scratch would have taken time and manpower we simply didn’t have.

So, when an AI startup approached us with a way to automate outbound efforts, it felt like perfect timing.

Instead of the generic cold blasts, the startup’s platform scraped social media pages to create emails that felt hand-crafted. How? We’re not exactly sure about the technical details, and that’s maybe a story for another day. But this was the promise: the more personalized the emails are, the more prospects find the offer relevant to them, the more likely they are to reply.

The rationale was sound, so we invested about US$1,600 as early adopters and handed over the keys.

But instead of giving us a reliable pipeline, it operated more like a black box.

The strategic danger of “outsourced learning”

The emails went out. And sure, we saw activity on the dashboard – the 45% open rates and 5% reply rate seemed great. The financial results though? A total flat line. There were some qualified opportunities, but zero revenue.

Here’s the part that really stung: We couldn’t even say why we weren’t getting the results we wanted.

Because we’d outsourced the execution to an external AI platform, we’d inadvertently outsourced our ability to learn. If we wanted to tweak the writing style and formatting, we had to email their engineers. If we wanted to change the reply flow, we had to talk to their account manager.

We couldn’t even verify the results because they were presented as is, without actual numbers on opens, clicks, and more. Plus, changes took more than a couple of business days. Failure without learning is just a sunk cost – for a decision-maker, that’s the most expensive kind of mistake.