Something has changed in what it means to be a senior associate. Not in title or compensation or the nominal description of the role. In the actual shape of the day.
A senior associate in a knowledge firm — consulting, law, accounting, advisory — used to spend a significant fraction of their day doing things that required their specific capabilities but not their deepest judgment. Drafting sections. Compiling analyses. Building the first version of something that would be revised up the chain. This was not wasted time; it was the mechanism by which they developed facility with the work. But it was not where the real value lived.
AI has changed what that work costs. The first draft, the compiled analysis, the structured summary — these are now available in minutes. Which means the senior associate’s day now consists of something different: deciding what to ask for, reviewing what comes back, and allocating the workflow.
Workflow allocation is a different skill
This sounds like a minor change. It is not. Workflow allocation requires a different kind of attention than first-draft production.
When you are drafting, the structure of the work itself guides you. You are inside the problem, following its shape. Mistakes surface naturally as you try to make things cohere.
When you are allocating — deciding what the AI should do, reviewing what it produced, deciding what to send back and why — you need to hold a clearer model of the expected output before you see it. You need to know what “good” looks like before you can evaluate whether you got it. This is a prior-knowledge problem, and it is harder than it sounds.
The senior associates who are thriving in this shift are the ones who had a strong mental model of the work to begin with. They always knew what a good memo looked like; they were just spending time writing it. Now they specify it and review it, which is faster and requires less production effort but more judgment from the start.
The ones struggling are the ones who were developing that mental model through the process of drafting. The drafting was how they learned what good looked like. AI has removed the scaffolding from their learning process.
What firms need to do about it
This is not an argument against AI-assisted work. It is an argument for thinking carefully about who is doing it and what they are actually developing through that work.
For senior associates who have already built strong mental models, the shift is a gift. They get their time back and can focus on higher-value judgment work. For junior-to-senior talent who are still in the development phase, the shift requires a deliberate replacement for what drafting used to teach.
The firms that recognize this are building explicit feedback loops: reviewing AI output with associates rather than just accepting it, doing structured critique sessions on AI-assisted work, preserving some first-draft production for developmental purposes even when AI could do it faster.
The firms that don’t are producing associates who are fast but hollow — able to allocate workflows they do not fully understand, toward outputs they cannot reliably evaluate.
That is a skills problem. It takes three to five years to become visible. And by then, it is expensive to fix.