About Poth

Building the future of feedback intelligence

Traditional surveys collect data. We create understanding.

The Problem

Millions spent. Nothing learned.

Organizations spend millions on employee surveys, customer feedback, and market research. Yet most of this data goes nowhere.

Why? Because traditional surveys ask everyone the same questions, generating shallow insights that don't reveal root causes. Analysis is manual, time-consuming, and biased by whoever interprets it.

The result: expensive data that sits in spreadsheets while real problems go unsolved.

Our Solution

Three innovations. One system.

Adaptive surveys that think

Every question is tailored to the respondent's previous answers. LLM-powered intelligence explores emerging themes in real-time.

A network, not a list

Responses connect to form a knowledge graph. AI agents navigate this network to find patterns and test hypotheses.

Insights, not spreadsheets

Statistically validated findings delivered automatically. No manual analysis required.

Technology

How we're different

LLM-Powered Adaptation

Each question is generated on the fly based on conversation context, allowing us to explore issues as deeply as needed without preset question trees.

Knowledge Graph Architecture

Responses aren't stored in rows and columns—they form a connected graph where relationships between answers reveal insights traditional analysis misses.

Hypothesis-Driven Agents

AI agents continuously form and test hypotheses about your organization, automatically collecting the data needed to validate or disprove them.

Statistical Rigor

We don't just spot correlations. Our system applies proper statistical testing to ensure findings are significant, not coincidental.

Principles

What we believe

Depth over breadth

We'd rather understand one problem completely than surface-scan a hundred issues. Our surveys dig deep.

Automation with transparency

AI should augment human decision-making, not replace it. We show our work and explain our reasoning.

Data privacy first

Your data is yours. We process it, analyze it, and help you understand it—but we never train models on it or sell it.