Granola Adds memory, and the Workflow Math Changes
What Cohere's latest move means for lead qualification in the year ahead.
For most of the last year, the conversation around RAG-as-a-service has been louder than the evidence. That is starting to change.
Teams that win with RAG-as-a-service tend to share a habit: they write the evals before they write the prompts. Everything else follows from that.
Eval harnesses, once an afterthought, are becoming the most important piece of code in many AI projects. Duolingo's team treats theirs the way an SRE team treats a runbook.
The cost curve matters here. Gemini 3 Pro is roughly an order of magnitude cheaper per token than the equivalent model 18 months ago, and that changes which problems are worth automating at all.
What Google DeepMind actually shipped with Mistral Large 3 is less a single capability and more a cluster of small, compounding improvements — the kind that only show up when you put a real workflow on top.
What Inflection actually shipped with Phi-4 is less a single capability and more a cluster of small, compounding improvements — the kind that only show up when you put a real workflow on top.
Inside Spotify, the rollout looked less like a moonshot and more like a slow migration. A pilot, a champion, a quiet expansion, a budget line.
The skeptical read is that we are watching a feature, not a platform. The optimistic read is that structured outputs is exactly the kind of feature that becomes a platform when nobody is paying attention.
None of this guarantees a clean story. Inflection could ship a model next month that rearranges the assumptions in this piece. But the direction of travel, for now, is clear enough to plan around.