We're in an AI slingshot moment, and the pullback isn't easy
The efficiency dip nobody in AI wants to talk about, and why the firms willing to endure it will win out.
Right now, I am slower at my job than I was last year.
I'm a product marketer. I know how to write. I know our customers, our product, our positioning. And yet I'm spending hours building processes in Claude for outputs that I could have completed myself in a third of the time, usually with a better first draft.
I'm telling you this because I think it's the honest and uncomfortable part of the AI conversation frequently being glossed over. We're all excited about the upside, and we should be. The potential is real. But so is the pullback. And if we don't talk about it honestly, we risk abandoning the process right before it pays off.
This is a slingshot moment. And slingshots only work if you're willing to pull back first.
The efficiency dip is real, and it's supposed to be there
When you genuinely rebuild how you work, not layer AI on top of broken processes but actually stop, examine, and reconstruct, you slow down before you speed up.
The mistake most people are making right now is thinking they can skip that part. They're adding bits of AI to the same inefficient, manual, human-patched workflows they've always had. This gives you a small boost. It's enough to feel like progress, but not enough to change anything fundamental.
You can't slingshot from a standing position. You have to pull back. The tension is what powers your momentum.
The rebuilding phase requires something uncomfortable: shining a light on everything that's broken. Asking hard questions. Where are we doing things manually that we don't need to? Where have we built workarounds on top of workarounds? Where is institutional habit masquerading as process? That kind of audit is slow, and it's humbling, and most organizations avoid it because the day-to-day doesn't stop while you're doing it.
But the firms, and the teams who are willing to do it? They're loading the slingshot.
Accounting has been here before
For those of us close to accounting firms, this moment has a familiar shape. Think about practice management software adoption. Workflow automation. The firms that invested in building those systems, really building them, not just buying a license and half-implementing it, had to slow down to do it. They had to migrate data, retrain staff, rethink how work moved through their practice. For a period of time, it was harder, not easier.

:format(avif))
:format(jpeg))

:format(jpeg))
:format(jpeg))
:format(jpeg))
:format(jpeg))
:format(jpeg))