A Post-AI Management Methodology

K·V·N

כּ · ו · נ

the root of kavanah — to aim, to align, to be ready

Three things every person being managed must carry into the work: the Know-How to do it, the Vision behind it, and the Negation that defines its edge.

#Why Now

Post-AI, the bottleneck
is alignment, not output.

When AI handles execution, the manager's job changes. Throughput is no longer the constraint. Judgment, intent, and taste are. The person being managed isn't a producer — they're an orchestrator of intelligence, human and machine.

AI faithfully executes ambiguity. Hand it a task with only positive direction and it will fill the unspoken space with its own priors — not the organization's. KVN forces the manager to be explicit on all three axes, because the AI-amplified worker will otherwise default to the model's defaults.

Kavanah — intention — is what separates ritual from rote. Post-AI, rote work gets automated. Only kavanah-laden work has human meaning, and only kavanah-laden management produces it.

#The Three Pillars

Three coordinates of intention

Each pillar names a distinct axis. Drop any one and the trajectory collapses into noise — well-executed, beautifully measured noise.

KKnow-How

Axis · Capacity

כֹּשֶׁרkosherfitness, capability, readiness

Not the ability to perform the task by hand — automation handles execution. The judgment to verify AI output, recognize when it's wrong, and know when to escalate.

Verification standards
How does the worker know an answer is right? What does 'good' look like?
Escalation triggers
When should they stop and ask? Naming the trigger is the manager's job, not the worker's.
Tool fluency
Which tools, models, and methods are sanctioned for this kind of work — and how each one fails.

VVision

Axis · Destination

כִּוּוּןkivundirection, orientation, aim

The end-state being pursued and the strategic context behind it. The 'why' that an AI cannot infer from the prompt alone.

Strategic context
Where this work sits in the longer arc. What it unblocks, what it signals, who it's for.
Success as end-state
Defined as the world-after, not the deliverable. 'Customers can self-serve refunds' beats 'ship a refund form.'
The single sentence
If the worker can't compress vision into one sentence, the manager hasn't transmitted it yet.

NNegation

Axis · Constraint

שְׁלִילָהshlilanegation, that which is excluded

What is explicitly out of scope, off-limits, or wrong-on-purpose — even when it appears to satisfy K and V. The negation space is where AI defaults silently win unless the manager preempts them.

Hard nos
Legal, ethical, brand red lines. Never crossed, even when it costs the deliverable.
Strategic nos
Technically valid solutions that violate strategy — e.g. 'don't optimize for speed at the cost of trust.'
False friends
Approaches that look right but aren't. The highest-value negations, because juniors and AI both fall into them.

#The Handoff

Mutual kavanah

Management under KVN is a contract. The manager transmits all three pillars; the worker transmits all three back. A delegation missing any of the six is delegating ambiguity.

Manager → Worker

The manager owes

K
Know-How
"Here's the method, the verification standard, and the escalation trigger."
V
Vision
"Here's the end-state we want and why it matters."
N
Negation
"Here's what would make a working answer still wrong."

Worker → Manager

The worker owes back

K
Demonstrated competence
Show the verification, not just the output.
V
Confirmed intent
Restate the vision in your own words before starting.
N
Surfaced near-misses
Name the approaches you considered and rejected — and why they sit in the negation space.

Template

Build X. Here’s the know-how you need or where to acquire it (K). Here’s the end-state and why we want it (V). Here’s what would make a working answer still wrong (N).

#Lineage

What the existing frameworks miss

Framework	K · Know-How	V · Vision	N · Negation
OKRs	—	✓	—
RACI	partial	—	—
Servant leadership	—	✓	—
Stoic management	—	—	partial
Constitutional AI	—	—	✓
KVN	✓	✓	✓

KVN doesn’t replace these — it forces a structural completeness they each leave open. The novelty is centering Negation as a coequal pillar, not a sidebar.

#Counter-questions

Pressure-tests

→Why three pillars, not five or seven?

Aiming requires three coordinates: capacity (what you can do), destination (where you're going), and constraint (the boundary that rules out misfires). Fewer fails to define a trajectory; more is decoration.

→Is this just a rebranded brief or RACI?

Briefs and RACI cover positive space — what to do, who owns it. KVN treats the negative space as coequal. In a post-AI workflow, omitting Negation isn't merely incomplete; it's the bug.

→Doesn't Vision imply its own negations?

Implicitly, yes. KVN insists on making them explicit. AI faithfully executes ambiguity by filling implicit negations with its own priors — not the organization's. What's implicit to a senior is invisible to a model.

→Does KVN apply to a task, a role, or a quarter?

All three. The framework is recursive across scales. A weekly standup transmits task-level KVN; a quarterly plan transmits role-level KVN; a strategy doc transmits org-level KVN.

Software built around
the same idea

Kavanah is a project platform built on the bet underneath KVN: that work fails not from lack of tools, but from lack of alignment between intent and action. It’s the operating layer for teams who want every task to carry kavanah.

Try Kavanah Read the book Discuss the methodology

A Post-AI Management Methodology

K·V·N

כּ · ו · נ

the root of kavanah — to aim, to align, to be ready

Three things every person being managed must carry into the work: the Know-How to do it, the Vision behind it, and the Negation that defines its edge.

#Why Now

Post-AI, the bottleneck
is alignment, not output.

Kavanah — intention — is what separates ritual from rote. Post-AI, rote work gets automated. Only kavanah-laden work has human meaning, and only kavanah-laden management produces it.

#The Three Pillars

Three coordinates of intention

Each pillar names a distinct axis. Drop any one and the trajectory collapses into noise — well-executed, beautifully measured noise.

KKnow-How

Axis · Capacity

כֹּשֶׁרkosherfitness, capability, readiness

Not the ability to perform the task by hand — automation handles execution. The judgment to verify AI output, recognize when it's wrong, and know when to escalate.

Verification standards
How does the worker know an answer is right? What does 'good' look like?
Escalation triggers
When should they stop and ask? Naming the trigger is the manager's job, not the worker's.
Tool fluency
Which tools, models, and methods are sanctioned for this kind of work — and how each one fails.

VVision

Axis · Destination

כִּוּוּןkivundirection, orientation, aim

The end-state being pursued and the strategic context behind it. The 'why' that an AI cannot infer from the prompt alone.

Strategic context
Where this work sits in the longer arc. What it unblocks, what it signals, who it's for.
Success as end-state
Defined as the world-after, not the deliverable. 'Customers can self-serve refunds' beats 'ship a refund form.'
The single sentence
If the worker can't compress vision into one sentence, the manager hasn't transmitted it yet.

NNegation

Axis · Constraint

שְׁלִילָהshlilanegation, that which is excluded

What is explicitly out of scope, off-limits, or wrong-on-purpose — even when it appears to satisfy K and V. The negation space is where AI defaults silently win unless the manager preempts them.

Hard nos
Legal, ethical, brand red lines. Never crossed, even when it costs the deliverable.
Strategic nos
Technically valid solutions that violate strategy — e.g. 'don't optimize for speed at the cost of trust.'
False friends
Approaches that look right but aren't. The highest-value negations, because juniors and AI both fall into them.

#The Handoff

Mutual kavanah

Management under KVN is a contract. The manager transmits all three pillars; the worker transmits all three back. A delegation missing any of the six is delegating ambiguity.

Manager → Worker

The manager owes

K
Know-How
"Here's the method, the verification standard, and the escalation trigger."
V
Vision
"Here's the end-state we want and why it matters."
N
Negation
"Here's what would make a working answer still wrong."

Worker → Manager

The worker owes back

K
Demonstrated competence
Show the verification, not just the output.
V
Confirmed intent
Restate the vision in your own words before starting.
N
Surfaced near-misses
Name the approaches you considered and rejected — and why they sit in the negation space.

Template

Build X. Here’s the know-how you need or where to acquire it (K). Here’s the end-state and why we want it (V). Here’s what would make a working answer still wrong (N).

#Lineage

What the existing frameworks miss

Framework	K · Know-How	V · Vision	N · Negation
OKRs	—	✓	—
RACI	partial	—	—
Servant leadership	—	✓	—
Stoic management	—	—	partial
Constitutional AI	—	—	✓
KVN	✓	✓	✓

KVN doesn’t replace these — it forces a structural completeness they each leave open. The novelty is centering Negation as a coequal pillar, not a sidebar.

#Counter-questions

Pressure-tests

→Why three pillars, not five or seven?

→Is this just a rebranded brief or RACI?

Briefs and RACI cover positive space — what to do, who owns it. KVN treats the negative space as coequal. In a post-AI workflow, omitting Negation isn't merely incomplete; it's the bug.

→Doesn't Vision imply its own negations?

→Does KVN apply to a task, a role, or a quarter?

All three. The framework is recursive across scales. A weekly standup transmits task-level KVN; a quarterly plan transmits role-level KVN; a strategy doc transmits org-level KVN.

Software built around
the same idea

Try Kavanah Read the book Discuss the methodology

K·V·N

Post-AI, the bottleneckis alignment, not output.

Three coordinates of intention

Mutual kavanah

The manager owes

The worker owes back

What the existing frameworks miss

Pressure-tests

Software built aroundthe same idea

K·V·N

Post-AI, the bottleneckis alignment, not output.

Three coordinates of intention

Mutual kavanah

The manager owes

The worker owes back

What the existing frameworks miss

Pressure-tests

Software built aroundthe same idea

Post-AI, the bottleneck
is alignment, not output.

Software built around
the same idea

Post-AI, the bottleneck
is alignment, not output.

Software built around
the same idea