The a16z Show · Friday, July 31, 2026
Decagon prioritizes in-house development for mission-critical aspects of model training, particularly when it's tightly coupled with their specific use cases like customer support. For more common tasks like data labeling, they prefer to leverage external vendors to accelerate model deployment.
“So, when we have, uh, open source, uh, uh, models that we want to fine tune, we find that if we can clearly tailor our eval, you know, to custom outcomes, that it's way better than just looking at loss curves over time.”
“So, for us, we're solving for customer support use cases and we've really honed in on that.”
“For other things like getting labeled data and measuring the diversity of our data sets, we're like, yep, these are tasks that are common across all companies. In which case, we want to buy things, uh, from other vendors because that'll just help us get those models production faster.”