01 / The starting point
The answers were spread across the business.
Visiting Media builds a visual sales platform for hospitality. Leadership wanted a clearer view of customer needs and the reasons some customers left, before committing to the next roadmap and organizational changes.
The evidence existed in recorded calls, support tickets, surveys, customer records, and a large codebase. Connecting it across those sources was a substantial research task. The company also wanted engineering and product teams to make more effective use of AI in their daily work.
02 / The first intervention
Read the evidence. Give leadership something it can act on.
Fleet of Geniuses built analysis pipelines to examine customer conversations, support issues, survey responses, and product code. We organized the findings into an evidence-backed view of opportunities and blockers, with customer language connected to the product issues behind it.
The assessment gave leadership a shared basis for setting priorities. The company used the findings in its roadmap and in organizing engineering, product, and UX work around the problems it wanted to solve.
It also established a practical starting point for a longer engagement: keep the customer evidence current, improve the data behind sales decisions, and help teams build with AI.
03 / What developed
Three connected capabilities inside the company.
The initial research developed into ongoing systems and working practices. Each connects to the same goal: make business knowledge useful and help the team act on it.
Customer Intelligence Platform
The platform brings together calls, support information, customer-health data, and a Salesforce mirror. Teams can prepare for customer conversations and investigate product questions using the underlying evidence. FOG architects and co-builds with Visiting Media’s in-house engineer, who is the primary author of the platform.
Hotel knowledge graph
Research pipelines connect properties to brands, owners, and operators, then reconcile those relationships with Salesforce. The work gives sales and RevOps a better basis for account research, contact sourcing, and record cleanup.
Engineering and product enablement
Working cohorts apply Claude Code to production tasks, supported by reusable skills, review practices, and shared conventions. Engineering, product, and UX staff work on the systems and decisions they already own.
04 / What changed
The client’s team is extending the capability.
The February engineering audit found a small AI skill library created by one contributor. By September, the repository contained 33 skills with eight contributors, and setup installed the shared tooling automatically. Claude review workflows and a review bot built by Visiting Media staff had become part of the development environment.
The intelligence platform tells a similar story: Visiting Media’s own engineer drives most of its development, with FOG contributing architecture and implementation. The client is actively shaping and extending its systems.
As of September 2026, the platform is in beta with initial users and working data ingestion. The next phase focuses on broader adoption: helping more teams bring shared customer evidence into their daily decisions.
Inside the systems
Evidence connected to its sources
The platform combines ingestion and analysis pipelines with account and product interfaces, including an agent accessible through Slack. FOG and the client team develop it together in Visiting Media’s environment.
Claude in engineering practice
Claude Code cohorts, shared skills, and AI review workflows support work in the production codebase. Review activity is evidence of changed practice, not a claim that human judgment has been removed.