Cassie Kozyrkov and decision intelligence — the reference
Cassie Kozyrkov is a statistician who served as Google's first Chief Decision Scientist from 2018 to 2023, where she trained over 20,000 employees in data-driven decision-making, and is now CEO of the AI advisory firm Kozyr. She frames decision intelligence as the discipline of turning information into better actions at scale, insisting the decision comes before the data: know what you would decide differently before you analyze. The term's foundational academic work is associated with Lorien Pratt; Kozyrkov operationalized and popularized the discipline at enterprise scale.
Cassie Kozyrkov & Decision Intelligence
Ask who made "decision intelligence" a household term in data circles and one name dominates: Cassie Kozyrkov, Google's first Chief Decision Scientist. This is the evergreen reference — who she is, what her framing actually says, what she did and did not originate, and what it means for how teams decide.
Last updated: 2026-08-20
TL;DR
Cassie Kozyrkov is a statistician who served as Google's first Chief Decision Scientist (2018–2023), training over 20,000 Google employees in data-driven decision-making and advising on hundreds of initiatives; since leaving Google she is CEO of Kozyr, an AI advisory firm working with senior leaders at organizations including NASA and Salesforce. She frames decision intelligence as turning information into better actions at scale — a blend of data science, behavioral science and managerial judgment in which the decision, not the data, comes first. The term's foundational academic work is associated with Lorien Pratt; Kozyrkov operationalized it inside a company and popularized it worldwide.
Who Is Cassie Kozyrkov?
A statistician by training, Kozyrkov joined Google in 2014 and in 2018 became its first Chief Decision Scientist — a role created around the observation that organizations were investing heavily in data while under-investing in the decision skills the data was supposed to serve. Her record in the role is the reason the title spread beyond Google:
- •Trained over 20,000 Google employees in data-driven decision-making through the company's internal programs
- •Advised on more than 500 initiatives, applying decision science across products and operations
- •Built "decision intelligence" into a named practice bridging statistics, machine learning and behavioral science
- •Left Google in 2023 to found Kozyr, an AI advisory firm, where she advises senior leaders at organizations including NASA and Salesforce
- •Writes one of the most-read bodies of public work on the discipline — her essays and talks are many practitioners' first contact with the term
What Does Her Framing Actually Say?
Stripped of buzzwords, three commitments recur across her published work — and each one is directly usable.
The decision comes before the data
Her signature move is to invert the usual order: instead of "we have data, what does it tell us?", ask what decision are we facing, and what would change our minds? Data earns its keep only when a decision would go differently because of it. Analysis without a decision attached is entertainment, however rigorous.
Decision intelligence blends disciplines on purpose
She describes the field as combining data science with the behavioral and social sciences of judgment — because real decisions fail as often on framing, incentives and psychology as on statistics. That blend is the discipline's identity, not an accident: the quantitative machinery serves human judgment, never the reverse.
Rigor belongs at the decision layer
Her enterprise work focused on making the process rigorous: who the decision-maker is, what the default action is, what evidence would justify switching, and how to test AI systems against those criteria before trusting them. It is decision hygiene at organizational scale — the same layer Gartner's DI definition formalizes.
Origin Credit — and What It Means for Teams
One distinction this reference exists to keep straight: Kozyrkov did not coin the concept in the academic sense — the foundational decision-intelligence work is associated with researcher Lorien Pratt, whose writing framed DI as a field linking decisions to outcomes. Kozyrkov's contribution is operational and cultural: she ran the discipline inside one of the world's largest companies, gave it a vocabulary practitioners actually use, and taught it at a scale no academic channel could reach. Both credits stand; they answer different questions. For teams, her framing translates directly:
Start from the decision
Before gathering evidence, write down the actual decision, the default, and what would change your mind — the framing step every structured deliberation on Argumentree begins with.
Make the reasoning inspectable
Her insistence that judgment stay visible alongside data maps to argument capture: claims, counter-arguments and evidence in the open, not compressed into a slide.
Separate deciding from wishing
The decision-first test — would anything change? — kills analysis theater. A rated argument tree makes the same test social: arguments that change no one's assessment reveal themselves.
Close the loop
Decision science without outcome review is a lecture. Decision records tied to later outcomes turn her training-room discipline into an institutional habit.
For the applied, essay-length treatment of these ideas in an enterprise context — including how they meet causal AI and large-scale operations — read our companion piece, Decision Intelligence: From Data to Action at Enterprise Scale. This page is the stable reference; the essay is the deep dive.
Explore further
What Is Decision Intelligence?
The hub: definition, lifecycle, decision rights and the four pillars.
Decision Intelligence Guide (Essay)
The applied enterprise treatment — the editorial companion to this reference.
Data-Driven Decision Making
The practice her decision-first test is designed to rescue from analysis theater.
Decision Intelligence for Teams
Where the discipline meets everyday team decisions.
DI vs Business Intelligence
The layer distinction her Google role was created to bridge.
What Is Decision Making?
The foundational hub on decision-making itself.
Frequently Asked Questions
Who is Cassie Kozyrkov?
A statistician who served as Google's first Chief Decision Scientist from 2018 to 2023, training over 20,000 employees in data-driven decision-making and advising on hundreds of initiatives. She is now CEO of Kozyr, an AI advisory firm, and one of the most widely read public voices on decision intelligence.
Did Cassie Kozyrkov invent decision intelligence?
Not in the academic sense. The foundational work on decision intelligence as a field is associated with researcher Lorien Pratt. Kozyrkov operationalized the discipline at Google, gave it its practitioner vocabulary, and popularized it at a scale that made the term mainstream — an operational and cultural contribution rather than an academic coinage.
Where is Cassie Kozyrkov now?
She left Google in 2023 and founded Kozyr, an AI advisory firm, where as CEO she advises senior leaders at large organizations on AI and decision-making. She continues to write and speak publicly on decision intelligence.
What is her definition of decision intelligence?
She frames decision intelligence as the discipline of turning information into better actions at scale — combining data science with behavioral science and managerial judgment, with the decision rather than the data as the starting point.
What is the decision-first principle?
Before analyzing anything, identify the decision you face, your default action, and what evidence would change your mind. If no realistic finding would alter the action, the analysis is not decision-support — a test that eliminates a large share of corporate analytics effort.
How does her framework relate to structured argumentation?
They meet at the decision layer. Her framing makes the decision explicit and keeps judgment visible alongside data; structured argumentation supplies the mechanism — claims, counter-arguments, evidence and ratings — that makes group judgment inspectable and improvable. Argumentree implements that mechanism.
References & Further Reading
Wikipedia — Cassie Kozyrkov
Biography: Google's first Chief Decision Scientist 2018–2023; 20,000+ employees trained; founding of Kozyr.
View source →Cassie Kozyrkov — Decision Intelligence (Substack)
Her current public writing on the discipline, in her own words.
View source →Cassie Kozyrkov — Medium
The essay archive that introduced many practitioners to decision intelligence.
View source →Cassie Kozyrkov — LinkedIn
Current role: CEO of Kozyr; previously Google's first Chief Decision Scientist.
View source →Put the Decision First
Her test is simple: know what would change your mind before you analyze. Argumentree gives your team the structure to run that test together — arguments, evidence and decisions, recorded.
Start Free Trial