Episode 88 · July 20, 2026

Debugging Human Communication with Andrea Goulet

Treating communication as technical infrastructure to prevent entropy in the AI-powered organization.

Overview

In this episode of the Inspire AI podcast, host visits with Andrea Goulet, founder of Debugging Human Communication, to explore why shared meaning is the ultimate bottleneck in automated systems. By applying Claude Shannon’s information theory to team dynamics, the discussion provides a framework for diagnosing noise, managing task conflict, and maintaining human oversight in agentic workflows. The conversation highlights how technical leaders must engineer communication to ensure that high-speed AI output remains aligned with strategic intent.

Why this matters

The primary risk in an AI-powered organization is not a lack of intelligence, but a failure of shared meaning as automation outpaces human synchronicity. Because entropy never stops, treating communication as debuggable infrastructure ensures that strategic goals are not lost to noise or misaligned definitions. Mastering these information dynamics is essential for leaders to maintain high-fidelity output as they scale complex, agentic systems.

Key takeaways

01Communication is organizational infrastructure that can be modeled using the source, channel, noise, receiver, and destination framework.
02Trust serves as a critical variable in message fidelity, determining how much noise is introduced or filtered during transmission.
03Healthy teams must distinguish between productive task conflict and destructive character attacks to leverage friction for better problem-solving.
04The 'fail fast' paradox illustrates how a single phrase can have divergent meanings across different roles and horizons without explicit alignment.
05In an agentic ecosystem, humans must remain the 'destination selectors' while agents serve as the vehicle and skills serve as the directions.
06Conway’s Law dictates that AI agent structures and software will inevitably mirror the communication patterns of the humans who build them.

Guests

Andrea Goulet

Communication Consultant & Coach; Keynote Speaker · Self-employed

FAQ

What is information entropy in organizational communication?

Information entropy is the natural loss of shared meaning as communication scales. In AI-powered organizations, entropy increases as automation outpaces human synchronicity, requiring leaders to treat communication as debuggable infrastructure.

How does Claude Shannon’s information theory apply to teams?

Shannon’s framework—source, channel, noise, receiver, and destination—can be used to diagnose communication breakdowns. By identifying where 'noise' enters the channel, teams can maintain higher message fidelity and trust.

What is the relationship between Conway’s Law and AI agents?

Conway’s Law states that software systems reflect the communication structures of the organization. In an agentic ecosystem, the architecture of AI agents will inevitably mirror the communication patterns of the humans building them.

What is the 'fail fast' paradox mentioned in the episode?

It occurs when a single phrase like 'fail fast' has divergent meanings across different roles. Without explicit alignment on definitions, one department may see it as innovation while another sees it as a lack of quality control.