A series by Emre Motan
What changes when analytical support depends less on a practitioner personally executing every investigation?
AI agents can help with investigation, reasoning, and synthesis. Turning that capacity into useful support also requires choices about how work enters the function, how evidence earns trust, what knowledge survives, and where people take responsibility.
This series develops a proposal for that operating model and examines what would make it reliable and sustainable. It is written for analytics leaders, practitioners, and founders thinking about how to build the function.
Start here
The AI-Native Analytics Organization
How could the organization change when shared execution lets practitioners support decisions beyond the investigations they personally carry out?
The reading guide
The AI-Native Analytics Organization
The organizational proposal: supported routes to answers, broader analytical ownership, reusable knowledge, and staffing for the work the function cannot sustainably cover.
What Product Analytics Is For, and What Changes When Agents Do the Work
The question underneath the proposal: why trustworthy numbers, accountability, and independence still need an owner when agents produce much of the analysis.
Questions the series explores
Work and capacity
What should practitioners delegate, how should they check it, and what actually limits the capacity of a function working with agents?
Trust and shared capability
How can evidence earn trust, corrections improve future work, and partners get useful answers through supported workflows?
People and implementation
How should teams add people, develop expertise, and build this capability out of real decisions and useful analytical work?
The guide will grow as essays are published. You can also browse all writing.