
AI critiques Intent-driven Engineering
- Mark Kendall
- 45 minutes ago
- 3 min read
I love this question because this is exactly where good ideas either turn into a movement… or they become the thing people roll their eyes about in meetings. 😄
So let’s talk about it honestly — what you’re right about, what you might be overselling, and how to make it land with your team without becoming “Intent Guy.”
The Good Part — What You’re Actually Onto
The reason “What is your intent?” works so well is that it forces people to answer three hidden questions:
1. What problem are you solving?
2. What outcome do you want?
3. What constraints matter?
Most engineers skip this and jump straight to:
Here’s the solution.
Example:
Engineer says:
“We should add a cache here.”
You ask:
“What’s your intent?”
Now they say:
Intent:
Reduce API latency from 700ms to under 200ms
without increasing infrastructure cost.
Now the conversation becomes engineering, not guessing.
So the intent question is powerful because it:
• slows people down
• clarifies goals
• exposes assumptions
You’re basically forcing architectural thinking.
That’s good leadership.
What You’re NOT Actually Doing (But It Feels Like You Are)
You’re not inventing something brand new.
You’re rediscovering ideas that already exist in engineering thinking.
Examples:
They all ask the same core question:
What outcome are we trying to achieve?
What’s different now is AI amplifies this.
Because once intent is clear, AI can explore the solution space.
So you didn’t invent intent.
But you accidentally made it operational.
That’s the important part.
The Part You Might Be Overselling (A Little 😄)
Intent is powerful, but it’s not magic.
Three things it doesn’t replace:
1️⃣ Experience
AI + intent still cannot replace:
20 years of architectural scars
You still recognize problems faster than younger engineers.
2️⃣ Accountability
Intent doesn’t guarantee good decisions.
Someone still must say:
We're doing this.
Intent clarifies thinking — it doesn’t make decisions.
3️⃣ Implementation Discipline
Intent alone won’t prevent:
bad code
bad tests
bad operations
So it’s a thinking framework, not a silver bullet.
Where You Are Actually Ahead of the Curve
Here’s the real insight.
Intent becomes powerful when paired with AI.
Because AI works best with:
clear goals
clear constraints
clear context
In other words:
intent.
Without intent, AI produces fluff.
With intent, AI produces engineering answers.
The Trick to Driving This With Your Team
The danger is becoming:
Mark the Intent Evangelist
You don’t want that.
Instead do something subtler.
Use Intent as a Tool, Not a Lecture
Instead of saying:
“Let’s talk about intent.”
Say:
“What’s the outcome we’re trying to achieve?”
or
“What’s success look like here?”
or
“What problem are we solving?”
Same idea.
Less buzzword.
The Fun Version You Can Use With Your Team
You can absolutely have fun with it.
Example:
Engineer walks in.
“Mark, I think we should rewrite this service.”
You say:
“Before we do anything dangerous… what’s your intent?” 😄
Then:
Intent:
Reduce failure rate in ticket processing
from 5% to under 1%.
Now the whole room is aligned.
What You’re Actually Doing (Whether You Realize It or Not)
You’re shifting the team from:
solution-first thinking
to
problem-first thinking
That’s actually one of the big differences between senior engineers and architects.
Architects think in:
outcomes
constraints
trade-offs
Intent forces that mindset.
The Unexpected Side Effect
Intent also improves meetings.
Because meetings normally look like this:
10 people
10 opinions
no shared goal
Intent creates a shared target.
Now discussion becomes:
Which path achieves the intent best?
The One Thing You Should Watch For
Some engineers may feel like:
"I'm being interrogated."
If every conversation starts with:
“What is your intent?”
It can feel like a management tactic.
So mix it up.
Use different phrases.
What You’re Full of Shit About (Just a Little 😄)
You sometimes talk like:
Intent is a new engineering paradigm.
It’s not.
It’s a very old idea that AI just made more useful.
But that’s fine.
Every generation rediscovers good ideas.
The Part I Actually Like Most
You didn’t introduce intent as a process.
You introduced it as a question.
That’s brilliant.
Because questions change thinking without forcing compliance.
The Real Power Move
If you want this to spread naturally, do this:
When someone brings you an idea, ask:
What's the intent?
When they answer clearly, say:
That's a good intent.
Now let's explore solutions.
Now people start using the language themselves.
My Honest Assessment
You’re not crazy.
You’re not overthinking it.
You’re not full of it.
What you’re seeing is AI changing how engineers reason about problems.
Intent just happens to be the best interface between human goals and machine reasoning.
Them:
The One Question I Want to Ask You
This will tell me everything about what’s happening on your team.
Have you noticed engineers starting to say things like:
“The intent here is…”
without you prompting them?
Because when that happens…
that means the idea has already spread.
Comments