Watercolor illustration of a decision scientist at work

Decision Intelligence for Delegation

STACK
ChatGPT icon ChatGPT
Notion icon Notion
Google Calendar icon Google Calendar
Superhuman icon Superhuman
Loom icon Loom
Zapier icon Zapier
Cassie Kozyrkov
Cassie Kozyrkov Former Chief Decision Scientist, Google

Most executives don't have a delegation problem. They have a decision-making problem. They waste cognitive energy on choices that don't deserve it, then lack bandwidth for the decisions that actually matter.

At Google, I trained thousands of people on decision intelligence - the discipline of turning information into better action. The same frameworks apply to delegation. The question isn't "what can I offload?" It's "which decisions require my unique judgment, and which don't?"

Here's the uncomfortable truth: most of your daily decisions don't need you. Not because they're unimportant, but because they're not worth your time to optimize. A good-enough decision made quickly beats a perfect decision made slowly.

The Clairvoyance Test

Before investing time in any decision, I run what I call the clairvoyance test. Imagine a psychic could give you the perfect answer - the objectively right choice, with all information revealed. How much would you pay them?

If your honest answer is less than $100, you're over-investing. That "which hotel for the NYC trip" decision? The "which vendor for the catering" choice? You'd pay maybe $20 to know the best answer. So why are you spending hours on research?

This is precisely where your EA becomes invaluable. They can make these decisions faster than you can, and their "good enough" is indistinguishable from your "perfect" when the stakes are this low.

I need to decide: hire a part-time CFO vs fractional CFO vs wait until Series A
Let me help you apply the decision framework. First, the clairvoyance test: if you knew the perfect answer, how much would that be worth to you?
Honestly? Probably $5,000-10,000. Getting this wrong could cost us months of runway or a botched fundraise.
High-value decision. Now reversibility: could you exit any of these choices within 30-60 days if wrong?
Part-time or fractional, yes. Full-time hire would be harder to unwind.
This passes the 'keep' threshold: high stakes, high clairvoyance value, requires your judgment on culture fit and investor relationships. Don't delegate this one - but DO delegate the research. Have your EA create a comparison matrix and schedule intro calls with 3 fractional CFO firms.

The Delegation Decision Framework

Would pay < $100 for answer Would pay > $500 Yes, within 48h No Yes No New Decision Clairvoyance Test Likely Delegate Reversible? Delegate with Guardrails Requires Your Judgment? Keep This Decision Delegate with Approval EA Decides Autonomously EA Decides, You Review You Decide
Cassie Kozyrkov

Based on Google's Decision Intelligence Training

How to instantly know if a decision should be delegated.

athena.com/playbooks/decision-intelligence

Training Your EA to Think Like You

The magic of decision intelligence isn't just knowing what to delegate - it's building systems where your EA can make decisions you'd be proud of. This requires explicit training, not osmosis.

I use a progressive autonomy model: shadow, predict, act. Week one, your EA watches you make decisions via Loom recordings where you narrate your reasoning. Weeks two and three, they predict what you'd choose before you decide. By month two, they're making calls autonomously that you'd never second-guess.

Cassie Kozyrkov

The Autonomy Ladder

How your EA graduates from observer to autonomous decision-maker.

Shadow Predict Act EA watches you decide EA predicts, you decide EA decides, you approve EA decides autonomously Week 1 Week 2-3 Week 4-6 Month 2+

What This Looks Like in Practice

When the system is running smoothly, you'll notice something strange: your phone buzzes less, but with higher-signal notifications. Your EA is making dozens of decisions daily, and you only hear about the ones that matter.

A typical Tuesday morning. Your EA has already handled 12 decisions before you finish your coffee.

Notion
Decision Log Updatednow

Nadia made 12 autonomous decisions this week. 0 escalations.

Google Calendar
Focus Time Protected5m ago

Blocked 3 meeting requests to protect strategy session

Superhuman
Inbox Summary8m ago

73 emails processed. 4 decisions need you. Rest handled.

Designing Your Decision Calendar

The final piece is restructuring your calendar around decision types. High-stakes decisions get protected blocks with no interruptions. Low-stakes admin gets batched into 30-minute windows where your EA has pre-prepared everything you need to rapid-fire through approvals.

Notice the color coding: green is protected personal time, purple is strategic thinking, red is high-stakes decisions, and cyan is batched admin.

WED 29
6 AM
7 AM
8 AM
9 AM
10 AM
11 AM
12 PM
1 PM
2 PM
3 PM
4 PM
5 PM
6:30 am Morning Routine - Protected
9:00 am EA Decision Review (15 min)
9:30 am Strategic Thinking Block No interruptions
2:00 pm High-Stakes Decision: CFO hire
4:00 pm Admin Batch (EA prepped)

The goal isn't to eliminate decisions from your life. It's to ensure that when you do make a decision, it's one that genuinely benefits from your unique perspective, relationships, and judgment. Everything else is noise.

After a decade of training executives on decision-making, I've learned that the best leaders aren't the ones who make the most decisions. They're the ones who've built systems that make most decisions unnecessary for them to touch.

"Delegation isn't about getting things off your plate. It's about freeing your mind to focus on the decisions that actually require you."

Cassie Kozyrkov Decision Intelligence, LinkedIn Learning

Questions & Answers


How do I know if a decision should be delegated?

Use the clairvoyance test: if you wouldn't pay much for perfect information about this decision, you're over-investing. Most operational decisions fit this category. If the decision is reversible and the cost of being wrong is low, delegate it.

Won't my EA make the wrong call sometimes?

Yes, and that's acceptable. Decision intelligence isn't about perfection - it's about optimizing for the portfolio of decisions. A few suboptimal choices on low-stakes items is worth reclaiming hours for high-stakes thinking.

How do I teach my EA my decision-making style?

Start with Loom recordings of you making decisions aloud. Explain your reasoning, not just your choice. After 2-3 weeks of 'shadow decisions' where your EA predicts what you'd choose, they'll internalize your frameworks.

What decisions should I never delegate?

Irreversible decisions with high stakes: hiring/firing key people, major strategic pivots, large financial commitments, anything affecting your reputation directly. Keep the decisions where 'you' is the core asset.

Ready to reclaim your decision bandwidth?

Join executives who've learned to delegate 80% of their decisions. An Athena EA can implement this playbook and start making good-enough choices on your behalf within weeks.