Page 47 — Torque revised from 25 Nm → 27 Nm following V2 retrofit (new bolt material, M8×40 stainless). Added an explicit calibrated torque-wrench verification step. Pages 48 & 51 reference the same torque value → updated automatically.
INDUSTRIAL AI · OVERVIEW
One platform.Your entire factory
Connect your sources as they are. One agent per source, all linked by a Knowledge Graph.
Good morning, Stefan.
VOLTA · BERLIN5 lines2,437 documents indexed01 MAY 2026 · 06:42
- 0Documents indexed
- 0Active users
- 0Messages this month
- 0QR scans this month
- 23 Apr
- 24 Apr
- 25 Apr
- 26 Apr
- 27 Apr
- 28 Apr
- 29 Apr
- 30 Apr
- 01 May
HIERARCHY
Your plant, mapped.
Plant → Line → Machine → Component. The hierarchy is editable, AI-generated from your connected sources, exportable as PNG or SVG. It maps every document, every work order, and every QR code to its equipment.
- 4+ levels deep
- AI generated from SharePoint · editable
- Auto tagging for bulk imports
DOCUMENT REPOSITORY
Every document, in the right place.
Tree, cards, or table: the view that works for you. Four analysis modes at upload (text, OCR, electrical diagrams, technical drawings). And when the AI search is unsure, it asks you a question instead of burying you in results that miss the point.
TEXT✦ Manual_
MV820_ v12_ FR.pdf Filler MV-820
ELECTRIC✦ Schema_
elec_ MV820_ rev_ B.dwg Filler MV-820
TEXT✦ Procedure_
Calibration_ V8.pdf Head 3
TEXT SOP_
changeover_ line04.docx Line 04
PNEUMATIC Pneumatic_
diagram_ capper.pdf Capper MV-700
OCR Retrofit_
report_ Q4_ 25.pdf Filler MV-820
TEXT Maintenance_
log_ 2026_ W17.xlsx Line 04
ELECTRIC Wiring_
diagram_ V12A.pdf V12-A sensor
TEXT Operator_
manual_ MV820_ DE.pdf Filler MV-820
TEXT Spare_
parts_ catalog_ 2026.pdf Filler MV-820
TEXT Pasteur_
setpoints_ v3.docx Pasteurizer
TEXT Bottling_
label_ specs.pdf Bottler B-200
- Tree / Cards / Table views
- Text · OCR · Electric · Pneumatic upserts
- AI search
- Versioning · AI diff
DOC DETAIL + CHAT
Ask the document in any language.
Have the document and it's conversation side by side. Our AI cites the exact passage it's anwers are based on. 15+ languages, electrical schematics, mechanical drawings. Multi-language voice on demand.
4.2 — Tightening procedure
Head assembly · Filler MV-820
- 10Verify bolt material before tightening.
- 11Clean threads with appropriate solvent.
- 12Tightening torque: 27 Nm (post-retrofit V2, M8×40 stainless).
- 13Tighten in cross-pattern, in 5 Nm increments.
- 14Verify with calibrated torque wrench.
- 15Mark the tightening with paint.
- 16Report in the WO log.
- Side by side view
- Every answer, sourced
- 15+ languages
- Electrical + mechanical schematics
MULTI-DOC CHAT
One question. Multiple documents.
Attach multiple documents to the same conversation. The agent cross-references the sources and answers with citations grouped per document, page by page. No more juggling 4 PDFs at once.
- Manual_MV820_v12_FR.pdf
- SOP_changeover_line04.docx
- Procedure_Calibration_V8.pdf
- Schema_elec_MV820_rev_B.dwg
How do we switch line 4 from product A to product B without losing torque calibration?
3 steps, cross-referencing all 4 documents:
1. Stop the line and lock out in CHANGEOVER mode (SOP §4.1).
2. Record the current torque 27 Nm on head 3 (Manual §4.2) before any disassembly.
3. Follow Calibration V8 §3 — target stays 27 Nm post-retrofit (M8×40 stainless, schematic fig. 4.2).
⚠ Calibrated torque-wrench validation required before restart.
- Cross-document synthesis
- Citations grouped per doc
- Thinking block · AI reasoning
- Click → jump to page
AI VERSION DIFF
Every retrofit, explained.
When a retrofit or a revision updates a document, you get a side-by-side comparison that shows exactly what changed, and an AI that explains it in plain language. Your teams know what to re-read, without hunting through the old version.
- V1
- V2
- V3
- V4
- V5
- 12Lines added
- 4Lines removed
- 3Lines changed
- 2Pages affected
- ·10Verify bolt material before tightening.
- ·11Clean threads with appropriate solvent.
- −12Tightening torque: 25 Nm (post-retrofit V1, M8×40 zinc).
- +12Tightening torque: 27 Nm (post-retrofit V2, M8×40 stainless).
- ·13Tighten in cross-pattern, in 5 Nm increments.
- +14Verify with calibrated torque wrench.
- ·15Mark the tightening with paint.
- AI summary · per page
- Line-by-line diff
- Diff PDF export
- Compare any 2 versions
QR + SHOP FLOOR
Scan. Ask.
Generate a QR code and link it to a document, a hierarchy level, a tag, or a specific piece of equipment. Technicians scan with any phone. They land on the right documents and ask their questions, wherever they are on the line.
- Document
- Hierarchy
- Tag
- Equipment
Filler MV-820 / Head 3
Align QR in the frame
The right doc, from the right QR.
- Document · Hierarchy · Tag · Equipment
- Scan count tracking
- PNG export
- Lasered onto metal
OEM / CLIENT PORTAL
Ship the intelligence with every machine.
You're a machine builder. Your equipment ships with paper manuals and a USB stick, and the field stays silent. Discover our new OEM portal: assign documents per machine, capture the questions operators ask when machines run or stop, and let the AI surface the recurring issues, by theme, by machine, by client.
- 4Active clients
- 10Machines deployed
- 121Documents shared
- 112Questions this month
Volta Beverages
- Filler MV-82012/14
- Capper MV-7009/11
- Pasteurizer P-80014/15
- Bottler B-20012/12
/01Calibration · torque47questions "Quel couple pour la tête 3 après retrofit V2 ?"
/02Error codes · ERR_4232questions "ERR_42_OVERTEMP — comment je résous ?"
/03Changeover · product A→B19questions "Procédure changeover Line 04, perdre la calibration ?"
/04Spare parts · M8 bolts14questions "Référence M8×40 inox compatible V2 ?"
- Client roster
- Doc-to-machine assignment
- AI questions clustering
- Headless embed API
AUTOMINSIGHT
Talk to your operational data.
When a machine throws an error, the agent links the PLC or SCADA fault to its fix, pulled straight from your technical documentation. And when a static dashboard can't show every metric, you just ask in plain language: the agent answers from your historian, or builds the chart on the fly.
"On overheat (ERR_42), check coolant flow. Step 1: open valve V12-A. See §4.2 page 124."
Manual_MV820_v12_FR.pdf · P. 124Main downtime cause on line 4 this month?
- 01
- 05
- 09
- 13
- 17
- 21
- 25
- 29
- 47%Jamming14.6 h
- 30%Changeover9.3 h
- 10%Maintenance3.1 h
- 13%Other4 h
- PLC live · OPC-UA · MQTT
- Fault-to-doc · closed loop
- OEE/TRS · charts on demand
- Scheduled reports via Email / Teams
KNOWLEDGE GRAPH
Your systems are blind to each other. The graph connects the dots.
The Knowledge Graph connects your machines, documents, work orders, production lots, and materials across every system. It grows richer over time, creating the semantic layer your systems were always missing.
Root cause: the supplier change introduced a +2 °C requirement on the pasteurizer that was never reflected in the SOP setpoint. Validation V07, which would have caught the drift, was skipped under operator pressure. No single system saw it — the graph did.
- Unify your systems
- Incremental · no rebuild
- Semantic + BM25 + traversal
PREDICTIVE MAINTENANCE
Predictive maintenance, connected to the systems you run.
AutomAssist brings industrial predictive maintenance within reach by linking your CMMS, SCADA supervision, ERP, MES and technical documents through a Knowledge Graph — one AI agent per source, read-only, operational in days. Instead of a mass data-normalisation project, the failure history and alarms you already record feed predictive and machine-learning models, alongside your preventive plans.
It runs with your existing maintenance software — a web-based, SaaS or on-premise CMMS — and extends it: enriched SCADA context, early-warning alerts, and machine records reachable on a phone or tablet at the point of the fault, without a heavy data-engineering programme.
- Connected, not migratedyour CMMS/GMAO keeps its role; AutomAssist links it to SCADA, ERP, MES and documents.
- Predictive from existing signalML fed by the Knowledge Graph, from data you already have — before any sensor rollout.
- Deploy your waycloud-agnostic or on-premise; web, smartphone and tablet access for the shop floor.
FAQ Platform & predictive maintenance.
- Predictive maintenance uses machine data and models to anticipate failures. AutomAssist connects your sources — SCADA supervision, CMMS, production history — through AI agents and a Knowledge Graph to deliver actionable predictions in days, with no long normalisation phase.
- Yes. AutomAssist integrates with existing maintenance software, whether SaaS, full-web or on-premise, and surfaces preventive and predictive recommendations on desktop, smartphone and tablet.
- No. With one agent per source and the Knowledge Graph, AutomAssist avoids heavy normalisation — you get results in days, not months.
AI dedicatedto the industry
Unify siloed systems. Find the right document in a quick QR code scan. Get reliable answers in a pinch.
- Microsoft for StartupsGenAI Studio · Station F · Paris
- Azure MarketplaceListed · co-sell activated · MACC eligible
- AI AmbassadorFrench Government · Osez l'IA