Decagon’s Playbook for Building Enterprise AI Applications

Decagon’s Playbook for Building Enterprise AI Applications – The a16z Show

Market views

Bullish

OPENSOURCE— Decagon shifted 90% of their AI workflow to open-source models for latency optimization and cost benefits. Fine-tuned smaller open-source models can outperform larger frontier models on specific tasks while being faster and cheaper.
AIMODELS— AI reasoning models have dramatically improved to handle complex meta-tasks like writing procedures, creating tests, and monitoring conversations that previously required significant human effort. This represents a major capability leap.
SAAS— Even with AGI, software and databases will remain necessary as agents need places to store work, pull information, and reason about things. CRMs and systems of record will continue to be valuable as the source of truth.
AIAGENTS— AI agents are creating more demand for services like customer support rather than reducing jobs, as companies realize the value of providing better support at lower costs and expand usage.

Not investment advice.