Internal AI Knowledge Assistant
Apollo
Apollo is an AI knowledge system serving 2,000+ frontline staff across 90 locations.
It answers operational questions in five languages, in seconds, with every response traced to one of 116 approved sources.
Apollo is an AI knowledge system serving 2,000+ frontline staff across 90 locations.
It answers operational questions in five languages, in seconds, with every response traced to one of 116 approved sources.
A store team member or manager who needed an operational answer originally had three options: submit a support ticket and wait, interrupt a manager, or search SOPs spread across several platforms.
Now the whole process is one message: tag Apollo, ask, and get an answer in seconds with its source linked
Multiplayer free-roam VR runs on an unusually complex stack: wireless streaming, full-body tracking, and the infrastructure behind it, with an equally intricate guest journey layered on top. Getting one answer mid-shift meant filing a ticket, pulling a manager off the floor, or combing through operational documents.
Every path was slow.
We realized we weren't just solving an efficiency problem; we were solving an engagement problem too.
Between the hardware, wireless tracking, and everything needed to keep a session running, staff already carried a lot. A tool with its own learning curve would have been one more thing to master, and the people who needed answers fastest would have been the least likely to reach for it. Apollo had to launch everywhere at once and ask nothing new.
"Designing a learning agent that could reach every team member, from floor staff mid-shift to new hires at onboarding, without adding another tool to learn was our biggest challenge. We 'failed fast' through several concepts
Why this failed: Create friction in existing workflow. Our team already juggles enough tools daily, one more would easily get forgotten.
Why this failed: Hard to access. Navigating through our internal knowledge base already was complicated enough. Apollo needed to live in a place with ease of access
Through successful development and testing, we were able to launch Apollo within Slack, letting team members use a tool they were already familiar with
Cut information retrieval time by 95%
Saved 120+ hours per week, organization-wide
Reduced new hire onboarding time by 25%, saving 50+ hours weekly
Improved quality of life for team members company-wide
Built a self-hosted tool to track and update information sources
Developed a stress-testing tool to keep Apollo's accuracy near 100%