I want to talk about something that doesn’t get said out loud much in freelance communities: AI made the productivity gains real, but I don’t have more free time than I did before.
Two years ago I was producing maybe 12 to 15 solid deliverables a month. That was my sustainable ceiling. Good work, reasonable turnaround, fair rates.
I started integrating AI tools. Drafting got faster. Research compilation got faster. First passes at strategy frameworks took twenty minutes instead of three hours. I absorbed the efficiency, charged the same rates, made significantly more per hour. That was the plan and it worked for about six months.
Then the slow adjustment happened. Clients noticed the turnaround. Expectations shifted. Then rates got renegotiated on the logic that if I can produce faster, why am I charging the same? One client straight up asked if I was using AI and when I said yes, they cut my rate by 30% while keeping the volume requirements the same.
I’m now producing closer to 25 deliverables a month at a net hourly rate that’s maybe 15% better than two years ago. Not the 3x improvement I thought I’d captured.
And this is just the freelance version. I watch the same thing happen in employed teams - AI gets adopted, headcount gets cut or frozen, the remaining people absorb the work, everyone’s “more productive” in output terms while feeling more pressured.
The technology captures real efficiency. But the efficiency gets extracted by whoever has pricing leverage in the relationship. Usually not the person doing the work. This isn’t specific to AI - it’s the same thing that happened with every productivity technology. But I didn’t expect it to happen this fast.
Curious whether others are actually managing to hold gains or whether the treadmill got them too.
Same pattern in SEO. Client budgets didn’t grow because AI tools exist. What happened instead is the benchmark for what counts as “enough work” shifted upward.
Previously a monthly retainer covered maybe eight pieces and a technical audit. Now the same budget is expected to cover twelve pieces because “you have AI, right?” The clients aren’t wrong that the cost per piece has dropped for me. But they extracted the entire margin. I’m doing more work for the same money and they’re paying less per unit.
The only protection I’ve found is anchoring pricing to outcomes rather than volume. If I charge for rankings improvement rather than content volume, AI helping me is actually a margin gain. But most clients want to pay for outputs they can count.
The in-house version of this is real. Our team adopted AI tools eighteen months ago. We didn’t backfill a role that opened up around the same time. Leadership framing was “you have AI now, you can absorb it.” Which we technically can. But absorbing more work with the same people while the expectation of quality stays constant isn’t the same as having more capacity.
What we actually have is more volume, same quality bar, same team size, and higher stress. The AI didn’t give us time back. It gave the company the option to not hire someone.
I made a deliberate decision not to let clients know how much AI I use. Not because I’m hiding it, but because the conversation always ends the same way.
What I charge is for judgment, not keystrokes. I know which brief is asking the wrong question. I know which angle will land with their specific audience. I know when the client’s instinct is bad and how to redirect it without making them feel bad. AI doesn’t do that. But if the client focuses on production speed they don’t see it.
So I protect the rate by keeping the focus on outcomes. That works until a client starts tracking hours or comparing rates with a freelancer who charges less because they pass the AI output through directly without the judgment layer. Then it’s a harder conversation.
The productivity trap is real and it’s not new, but the speed of the current cycle is.
When email replaced paper correspondence, productivity expectations adjusted, but over decades. When digital tools replaced analog workflows, same thing. The adjustment was slow enough that it felt organic. AI is compressing that cycle dramatically. What usually takes ten years to recalibrate is happening in eighteen months.
From a management perspective I’d say the error is measuring AI impact through volume rather than value. If your team is using AI to produce more of the same, you’ve just run faster on the treadmill. If they’re using it to do qualitatively different work - deeper analysis, faster testing of more hypotheses, strategic work that previously didn’t fit in the calendar - that’s a real gain. Most organizations aren’t structured to capture the second category.
This makes me think about school in a way I hadn’t before.
Teachers give more assignments now partly because AI can help students finish them faster. So the homework load went up, not down. That’s the exact same thing you’re describing. Technology makes the work go faster, and the response is more work rather than less.
I thought AI was going to give me more time for the stuff I actually want to do. Instead I have more assignments. Same trap, just smaller scale.