Is it worth it to launch a new feature internally when you know its not fully ready yet?

A few weeks ago, our team went offsite to a beautiful Airbnb in Marin to do a last round of debugging before we launch Agents - probably our biggest feature launch to date, coming in early September.

Our time in Marin mostly consisted of 8.5 engineers staring at our computers for about 80 hours straight, but ultimately failing to get to all of the bugs we were hoping to squash.

We wanted to have a solid month of testing before taking Agents live for our clients, which left us with a decision to make:

A) launch internally when we knew the Agents feature wasn't fully ready yet

B) take it live externally without the full month of testing
C) push back our external launch

We opted for option A. I'm glad we did - our first week's results were strong, with 13 leads rescued from being errantly dumped into the unqualified bucket.

One of the leads was a property manager who visited our site a total of 4 times, including two visits to our blog on how a unified inbox simplifies real estate operations with AI.

When she booked a demo on her 4th visit, she selected 1 - 10 people as the company size - which would have resulted in the lead being discarded prior to our internal Agents implementation. But instead, our Leads Agent called her with the intent of qualifying her (confirming our min pricing was something they were open to).

The Agent called with a full understanding of who she was, who her company was, and the exact journey she had taken on our website across all 4 of her visits. During the call, it was tasked with learning more about what her company does and why she were interested in Conduit - after which it booked her for a demo right then and there on the call.

We're excited to see our vision for agents manifest itself internally, but there was definitely a level of risk associated with putting them into production knowing they weren't fully ready.

If put in a similar situation, what would your team have done?