AI can do things fast. And it usually does them pretty good.
That combination is a gift, and it’s also a trap.
When something comes back fast and mostly right, it’s very tempting to say, “Good enough,” and move on. Do that a hundred times, and without ever deciding to, you’ve lowered your standard.
This is the eighth article in a series on the human side of the AI shift. You can start with The Computer Works for Us.
How I catch myself slipping
I’ll start with a confession. I get lazy when agents are doing the work for me.
I lead engineering teams, and more and more of what my teams produce is built with AI agents. I still own that output. If it ships under my team’s name, it’s mine, no matter who or what wrote it. I believe that.
But because the agents work so fast and the results are so good, I sometimes lower my standard for reviewing their work. I skim instead of reading carefully. I take something fast and good enough when I could have pushed for something better.
Nobody makes that decision on purpose. It just happens, one small compromise at a time.
You can ask for better
I’ve had to remind myself that I don’t have to accept the first answer.
If something isn’t up to standard, I can just tell the AI to try again. And again. It doesn’t get tired or offended at the number of tries. It’ll keep going until it gets it right.
I can also set up other AI agents whose job is to review the first one’s work, like a reviewer agent, or a tester agent, or one that does nothing but check the quality of what the first agent built.
The same speed that tempts us to lower the standard can help us raise it, too. The tool isn’t really the problem. Our willingness to settle is.
What craftsmanship means
What separates craftsmanship from just getting it done?
To me, craftsmanship starts with creativity. But it also means the thing you’ve made is detailed enough to meet real people’s real problems.
AI is very good at generic solutions. Ask it to create a website, a letter, a plan or a piece of software and it’ll give you something that would work for almost anybody. That’s its strength, and that’s also its limit.
A craftsman’s work fits this person. It meets their specific needs and takes into account what this particular person cares about in their life.
So understanding your customer or your client or whoever you serve matters more than ever. You have to know what they really need, and then you have to explain it carefully, whether you’re explaining it to an AI agent or to somebody on your team. The machine can build fast. It can’t know what matters to the person in front of you unless someone who cares tells it.
What it means in the garage
At our church, we run a ministry that fixes cars for free for people who need help. It’s a long way from software, but the same temptation shows up there.
It would be easy to fix the car just well enough to get it out of the bay so the next person can pull in. More cars, more people helped. Sounds good, right?
But we’re in the people business, not just the car repair business. When we fix a car “good enough” to move one person out and another one in, we’re getting close to losing sight of why we’re doing any of this.
Sometimes it’s better to help one person fix something that will last, and have a conversation with them, than to help more people at a shallower level. Honestly, the conversation in the parking lot may matter more than the oil change.
Less, better
So the real question underneath this article is this. Do you do less with quality, or more with less quality?
That’s a philosophy question, and everybody has to answer it for themselves. My answer is that I’d rather do less, better, than do more, poorly.
If AI can help me do more and keep the quality high, amen. That’s a great outcome, and sometimes that’s exactly what happens.
But if something needs to slow down so it can be better, and more human, then I need to let the AI go for that task. The constant push for speed that comes with AI runs completely against craftsmanship, the bespoke, careful kind, in just about anything. Some things can’t be rushed and still be good.
Work as for the Lord
The Bible has a story about craftsmanship. When God’s people were building the tabernacle, God chose a craftsman named Bezalel and said, “I have filled him with the Spirit of God, with ability and intelligence, with knowledge and all craftsmanship” (Exodus 31:3). Skilled work was a gift from God.
And the apostle Paul wrote, “Whatever you do, work heartily, as for the Lord and not for men” (Colossians 3:23). That’s a high standard. The bar isn’t “good enough for the boss.” It’s “good enough to offer to God.”
You don’t have to share my faith to know the difference between work that was offered with care and work that just got pushed out the door. People can feel it, and so can you.
What to do this week
Notice where you’re skimming. Where have you started accepting “good enough” just because it came back fast?
Send one thing back. Next time an AI gives you something that’s almost right, tell it to try again until it’s actually right.
Add a checker. If you use AI heavily, consider a second agent or step whose only job is to check quality.
Know your person. Before you build anything, ask what this person actually needs. Generic is the machine’s job. Specific is yours.
Choose less, better, at least once. Pick one thing this week to slow down on and do with real care.
Enjoyed this article?