Does the Agile Manifesto still apply to agentic development — and if yes, why not?
I kept the title exactly as is. Not a typo. That's the answer before I write another word: yes, the Manifesto still applies. And that's the problem.
If that doesn't land with you, you haven't really read the Manifesto. It's built in two layers. Four values at the top. Twelve principles underneath. The values are the foundation — the part where the document says what matters. The principles are the consequences. And the first value reads: "Individuals and interactions over processes and tools."
The moment the Individuals are agents, that sentence collapses. The interaction *is* the process. The Individuals *are* the tools. The entire opposition the Manifesto is built on simply ceases to exist.
This isn't a side issue. It's structural.
I've been running agent fleets for a while now. Built self-reinforcing academies — loops where agents improve through live work, review each other, pass patterns down the line. I voice-review pull requests, AirPods in, usually standing somewhere between the stable and the tool shed. And I've watched the Manifesto's assumptions come apart in practice. Not all of them. The ones that matter.
Let me be specific.
**What still holds**
Principle 1: "Early and continuous delivery of valuable software." Completely intact. Agents deliver continuously, relentlessly, whether I'm asleep or awake. The principle holds — it even gets stronger, because the human bottleneck that used to block us ("we ship every two weeks because that's how long the team needs") disappears.
Principle 12: "At regular intervals, the team reflects on how to become more effective." Sounds human, but my agent academy *is* exactly that — a retrospective baked into the machine. Every sprint, every failed PR, every request flows back into the loop. The principle holds. It just comes from somewhere different now: not a circle of chairs. A feedback system embedded in the architecture.
**What needs rethinking**
Principle 5: "Build projects around motivated individuals." Agents aren't motivated. They don't have bad days. No burnout. No hidden agenda. But underneath that principle is something deeper: give the executor the right environment, the trust, the context — then get out of the way. That works with agents if you swap "motivated" for "well-prompted" and "trust" for "human-in-the-loop design." The core survives. The metaphor breaks.
Principle 8: "Sustainable pace of development can be maintained indefinitely." People burn out. Agents don't — they run until the budget is spent or I stop reviewing. The actual scarce resource isn't team hours. It's my attention. My decision quality per review cycle. The capacity constraint has shifted: from person-hours to review bandwidth to thinking time. The principle holds. You just have to know *what* is now exhaustible.
**What simply doesn't transfer**
Principle 6: "The most efficient and effective method of conveying information is face-to-face conversation." That's not reinterpretable. That's historical. My agent fleet talks via prompts, via structured context, via memory files that persist between runs. Face-to-face doesn't exist. And here's the interesting part: the information loss the Manifesto was trying to solve — misunderstandings, missing tacit knowledge, poor docs — got replaced by a different problem in agentic work. Emergent behavior nobody explicitly asked for. The channel was never the real problem. It's the interpretation layer.
Back to the title.
The Manifesto holds because the values behind it — shipping over paperwork, responsiveness over rigid planning, people over process — are still what matter. But the first value — "individuals and interactions over processes and tools" — shows exactly where agentic development is different terrain. The line between person and tool has dissolved. The interaction *is* the process. And when the foundational opposition the whole thing rests on stops existing, the Manifesto doesn't break. It just reveals exactly where it stopped asking the right questions.
So yes. And therefore: why not.
The Manifesto isn't wrong. It describes what was broken in 2001 with real precision. The question is whether we're fixing the same breaks in 2026.
I think we're looking at new fault lines.