The knowledge retiring technicians carry isn't in any manual.
Lore is a voice AI mentor that captures the tacit expertise of senior technicians and delivers it to junior technicians on the shop floor. Hands dirty, no screen. Just ask.
Built at the Activate Your Voice Hackathon — Speechmatics × The AI Collective Paris Track 1: Communication & Human Experience · February 28 – March 1, 2026
By 2033, 1 in 5 aviation maintenance jobs will go unfilled. Every retiring senior technician takes decades of contextual knowledge that was never written down — the exceptions, the quirks of specific airframes, the patterns that manuals can't capture.
Existing tools (Zymbly, LexX, AWS Q) do RAG on explicit documents: manuals, SOPs, service bulletins. None capture what seniors never thought to write down.
Lore operates in three modes:
After an intervention, the senior debriefs Lore by voice. Lore actively interviews — asks follow-up questions, flags ambiguities, and stores contextual knowledge linked to the specific airframe, component, and conditions.
A junior technician, hands in the machine, asks a question by voice. Lore responds with:
- Layer 1 — Relevant SOP/manual excerpt (RAG on official docs)
- Layer 2 — Contextual wisdom from senior's oral knowledge base
- Layer 3 — This specific aircraft's maintenance history
Voice-first intervention logging: "Lore, log: N1 vibration 2.4 units, temperature 6°C, no escalation." Adds to the aircraft's persistent memory.
Rule #1: Lore never contradicts a SOP. It completes it.
Scene: Hangar, 23h. Thomas (junior, 2 yrs experience) alone on Airbus A320, tail F-GKXA, CFM56-5B engine. He finds an anomaly not in the job card.
Thomas: "Lore, I'm on F-GKXA, CFM56-5B, fan section. I'm getting an unusual
low-frequency vibration on the N1 shaft. It's not in the job card.
What do I know about this?"
Lore: "According to SOP 72-21-00, N1 vibration above 4 units requires
escalation. However, Marc noted in October that F-GKXA specifically
shows a harmonic resonance between 2-3 units in cold conditions —
below 8°C. It's a known characteristic of this airframe, not a defect.
He recommended logging it and monitoring across the next two cycles
before escalating."
Thomas: "Got it. Logging: vibration at 2.4 units, temperature 6°C, no
escalation per Marc's note. Lore, add this to F-GKXA's memory."
Lore: "Logged. F-GKXA memory updated."
| Layer | Technology |
|---|---|
| Voice Input | Speechmatics (real-time STT, noise-robust) |
| LLM | OpenAI GPT-4o |
| Memory + RAG | Backboard (threads + document retrieval) |
| Voice Output | OpenAI TTS (gpt-4o-mini-tts) |
| Frontend | Next.js 14, TypeScript, Tailwind CSS |
| Deploy | Vercel |
- RAG retrieves what you put in. Lore extracts what seniors never thought to write down — through active dialogue, not passive ingestion.
- Knowledge graph relationships between airframe × component × condition × expert
- Contradiction detection: if Marc and Jean-Pierre disagree, Lore flags it rather than averaging
- Confidence scoring: observations confirmed by 4 technicians over 3 years outweigh isolated notes
Built in 24 hours by a team of 4.
# Install dependencies
npm install
# Configure environment
cp frontend/.env.example frontend/.env.local
# Fill in API keys in frontend/.env.local
# Create/validate Backboard assistant + threads
npm run setup-backboard
# Optional: seed demo memory
npm run seed-backboard
# Run frontend dev server from root
npm run devOpen http://localhost:3000.