Discovery Cycles

A Discovery Cycle is the fundamental unit of work in Horizon. It is a structured process where AI interviews employees across your organization through one-to-one conversations, extracts the relevant information, and turns it into Insights and then into actionable Initiatives. All the value Horizon creates starts here.

How does it work?

• Phase 1: Data collection. The admin uploads company information: organizational structure, strategic goals, focus areas, and the list of employees to interview. Horizon needs context to ask relevant questions, not generic ones.

• Phase 2: Interview preparation. Horizon automatically generates an APQC-style matrix for each business unit: expected processes, subprocesses, and activities. On top of that, it creates an individualized interview plan for each employee based on role, area, and cycle goals. It is the equivalent of doing interview research beforehand, but at organizational scale.

• Phase 3: Review. The admin reviews the generated plans and can adjust questions, priorities, or focus areas before launch. A superuser gives final approval. Nothing runs without human validation.

• Phase 4: Interviews. AI runs each conversation using the full accumulated context: company data, APQC matrix, individual plan, and strategic goals. Collaborators can speak or write, and the AI follows up on what matters most. This is not a survey, it is an advisory conversation tailored to each person.

• Deep Dives: focused discovery. When an Insight needs more evidence or validation, you launch a Deep Dive: a focused interview cycle around that specific Insight, with affected employees and targeted questions to validate feasibility, measure real impact, and build the business case for the Initiative.

Why does it matter?

Discovery Cycles compress into weeks what traditional consulting takes months to complete. Speed matters, but depth at scale is the bigger advantage. A consulting firm interviews 20 people and extrapolates. Horizon interviews everyone with individualized context and produces findings traceable to the source. Every Insight has evidence, every recommendation has data. No black box.