01 · SENSOR
Measured signal
Temperature, gravity, current, position, door state.
AEDDIX Systems GmbH — Wattens, Tirol, Austria
Industrial AIoT
Machines report.Agents triage.People act.
AEDDIX turns mixed physical estates into a governed system of action. Sensors report. Agents correlate, predict, and triage. People intervene only where policy requires — while rehearsal on a copy prepares the live estate for the bad day.
Trusted in production
Programs
Cloud speedOn-prem controlData sovereignty
AIoT modeling · agentic simulation · bounded physical action
01 — The pipeline
See everything. Act through one gate.
Signals from any vendor cross the trust boundary into one living estate model. Agents and the simulator may only propose. Nothing reaches a person or a machine except through one customer-owned policy gate, and every hop is written to the ledger. Pick an estate and watch the run; every node states whether it is live, in productisation, or roadmap.
ONE PIPELINE · CHOOSE THE ESTATE IT RUNS ON
CLICK A NODE TO INSPECT · DRAG TO REARRANGE · DOUBLE-CLICK TO FOCUSSTANDING BY
NOW
One pipeline. One gate. Evidence by default.
WHAT IT DOES
Signals from any vendor land in one living estate model. Agents and the simulator may only propose. Nothing reaches a person or a machine except through one customer-owned policy gate.
GUARANTEE
SEE EVERYTHING · ACT THROUGH ONE GATE
EVIDENCE
Live, productising, and roadmap stages labelled on every node
LIVE TODAYIN PRODUCTISATION · GATE 6–12 MOROADMAP · GATES 12–24 MO
REFERENCE DESIGN · ROADMAP TARGETS, NOT SHIPPED CLAIMS
Compounding context
The estate graph deepens with every asset and every resolved event.
Deterministic replay
Any incident re-run bit for bit; the response drilled before it is real.
One gate, no side doors
Nothing acts except through policy — agent, human, or external system.
Evidence by default
Who · what · why · policy version, written on every hop.
Why this compounds. Incumbents bolt AI onto dashboards. AEDDIX built the pipeline — intelligence touches the world only through the gate.
Fig. 01 · Reference design · public operating model
Illustrative events · no platform internals shown
02 — Operating model
One estate model. Two operating loops. One policy.
The same ladder works in a brewery, a pharmacy, and a fleet. The live loop handles today's exception; the rehearsal loop tests tomorrow's response on a copy.
CONTEXT COMPOUNDS AS THE SIGNAL CLIMBS
02 · GROUP
Correlated condition
Several signals agree; noise becomes evidence.
03 · PROCESS
Predicted operational risk
Batch, route, or cold-room tolerances add stakes and time.
04 · MACHINE
Owned physical outcome
One bounded action, named owner, deadline, and audit trail.
LIVE TODAY
Live loop
Machines report → agents triage → people act. Production evidence: 275 live devices, 365 recorded decisions, and 58 issued invoices.
IN PRODUCTISATION · GATE 6–12 MO
Rehearsal loop
Brief → run → judge → replay → promote. Agents drill the bad day on a deterministic copy; the live estate never feels the experiment.
Gated roadmap
Earn the right to move from advice to bounded physical action.
Software first, rehearsal second, edge third, physical AI last. Hardware enters only where it lowers deployment friction, protects sovereignty, or closes a proven operational loop. Each phase has a gate; no phase starts on faith.
01 · 0–6 MO · UNDERWAY
Productise the field patterns
Package cold-chain, fleet, and process workflows into deployment templates; formalise controls, evaluation, and rollout.
GATE · 3 DEPLOYMENT TEMPLATES SHIPPED
02 · 6–12 MO · IN PRODUCTISATION
Ship the rehearsal loop
Governed drills, operating tolerances, and policy comparison over the deployed estates. The simulator becomes a product surface.
GATE · REHEARSAL LOOP GA
03 · 12–18 MO · ROADMAP
Edge runner and hardware pilot
An on-prem model runner with store-and-forward continuity; reference hardware only where commodity devices leave a gap.
GATE · ONE EDGE PILOT LIVE IN THE FIELD
04 · 18–24 MO · ROADMAP
Supervised physical AI
Bounded actuation under customer policy, human approval, fail-safe states, and a full evidence trail. The only path back to the machines.
GATE · ACTUATION UNDER POLICY
Planning targets, not shipped claims. No physical action ships that has not been rehearsed in simulation under the exact policy that will govern it live.
02-B — Control plane
Govern the model. Rehearse the change. Keep people in command.
THE CONTROL PLANE
01 · MODEL
One estate model for heterogeneous everything
Sensors become groups, groups become processes, and processes belong to machines and environments. Context compounds without replacing the systems already running the operation.
02 · REHEARSE
Run the bad day against a copy first
Drill a failed cold room, a weakening reefer, or a drifting batch against a sandboxed estate. Compare responses and promote only the policy that survives the run.
03 · GOVERN & ACT
Only the exception seeks a human
Agents correlate, predict, and triage under customer-owned policy, and nothing acts except through one gate. Routine events close with a recorded verdict; the exception becomes one task with an owner, reason, and deadline.
AIoT, made real
The model is a guest in the customer's house.
It sees what policy shows, says only what the record supports, and can be replaced without rebuilding the operation.
AT THE ASSET
AIoT starts with physical context
Telemetry is attached to a real machine, vehicle, room, batch, route, and operating state—not handed to a model as an anonymous row.
AT THE BOUNDARY
On-prem AI when the data stays on site
Models and rules can run in your cloud, on-premises, or near the edge when sovereignty, latency, or continuity requires it.
AT THE MODEL
Closed scope, grounded explanation
Models operate against approved vocabulary, procedures, and decision examples. Every figure must resolve to the record; missing data stays missing.
AT THE REHEARSAL
Models earn promotion in simulation
A model or policy change runs against customer-authored tolerances in a sandbox before it can enter the live operating loop.
03 — Problem & timing
When everything alarms, nothing does. Industrial SMEs need protected attention, not another dashboard.
The operating model has to change now: experienced operators are scarce, traceability obligations are rising, and AI can finally work near the physical process under policy.
PRESSURE 01
Signal overwhelms attention
Ten sensors are manageable. A few sites and a hundred sensors turn the wall of red into background noise.
PRESSURE 02
Knowledge sits outside the process
What a reading means still lives in a veteran's head, an inbox, or a spreadsheet instead of the asset record.
PRESSURE 03
Evidence requirements rise
Pharma, food, logistics, and critical operations must show what happened, who decided, and under which policy.
PRESSURE 04
Simulation and physical AI matured
Models can now rehearse near sovereign data, predict the bad day, and participate in bounded physical workflows.
Market window
Growing, fragmented, time-sensitive.
- $17.5B
- AIoT market in 2026
- $38B
- net-new annual AIoT market by 2030
- $8.5B
- cold-chain monitoring wedge in 2026
- 2026–29
- founder capture-window thesis
Market estimates: The Business Research Company, 2026. The capture window is AEDDIX's thesis, not an analyst forecast.
03-B — Business case
One platform. Two reasons to care.
FOR THE OPERATOR
What an SME buys
A practical control layer over the systems already running the business. Start with one expensive workflow; expand without replacing the source.
- 01
No rip-and-replace. Keep existing machines, vendors, and databases. AEDDIX works across them.
- 02
One accountable queue. Alarms, evidence, decisions, owners, and deadlines meet in one workflow.
- 03
Human control. Agents classify and draft. Your people approve, edit, or dismiss.
- 04
Audit built in. The path from field signal to human decision remains traceable.
FOR THE COMPANY
Why the model scales
The first deployment solves a vertical problem. The underlying orchestration, governance, and work engine is horizontal.
- 01
Vertical wedge, horizontal platform. Cold chain, telematics, and process operations reuse one control plane.
- 02
Software attached to operations. Recurring workflows and connected assets support recurring customer relationships.
- 03
Decisions compound. Each resolved event can become tested, versioned policy instead of tribal knowledge.
- 04
Deployments become templates. Connectors, asset models, rules, and operator surfaces are reused across customers.
04 — Records
Every asset gets a record. Every alarm gets a verdict. Every task gets an owner.
Customer relationship management taught businesses to never lose track of a person. AEDDIX applies the same discipline to your operational assets: a fermenter, a delivery van, and a pharmacy fridge each carry their history, their context, and their open items — and the system keeps the relationship warm so your team doesn't have to.
- WHO IT IS
- The record. Identity, location, lineage, and live state for every asset — however unlike its neighbours it is.
- WHAT HAPPENED
- The history. Every excursion, verdict, and intervention, kept in order and kept auditable.
- WHAT'S NEXT
- The next action. One owner, one clearly-worded task, one deadline. Suggested by an agent, decided by a human.
ASSETReefer van · RV-07STATUSen route · last seen 41 s ago ●OPEN ITEMS1
HISTORY
17:02 · door open 4 min at stop
└ verdict: benign — agent, confidence 0.87
16:40 · cargo temp drift +0.8 °C
└ task created, owner assigned
NEXT ACTION
Verify door seal at depot — due today 18:00
05 — Deployed
Where it runs today.
Protect the payload, not the dashboard: medicine, batches, deliveries, uptime, and the operator attention required to keep them moving.
01
Brewing & beverage
Fermentation and taproom operations
02
Cold chain & pharma
Temperature integrity with full traceability
03
Logistics & telematics
Fleet behaviour beside cargo condition
04
Gastronomy
Digital taproom and live inventory
05
Consulting & DIY
Competence for hire, open frameworks to build on
06 — Trust
Compliance is an operating discipline—not a row of unearned badges.
AEDDIX maps each deployment to the privacy, security, and supply-chain obligations that actually apply. The control baseline is designed to support regulated workloads while independent attestations and certifications remain explicit roadmap items.
Current posture: framework mapping and assurance roadmap. AEDDIX does not claim a SOC 2 report, ISO certification, or generic HIPAA certification on this page.
HIPAA
DEPLOYMENT-SPECIFICUNITED STATES · HEALTH INFORMATION
There is no official HIPAA certification. A US healthcare deployment requires the applicable administrative, physical, and technical safeguards plus a Business Associate Agreement when AEDDIX handles PHI.
SOC 2 Type II
ASSURANCE ROADMAPSERVICE-ORGANIZATION ASSURANCE
Target an independent examination of control operating effectiveness over time, beginning with security and extending to relevant availability, confidentiality, and privacy criteria.
Québec Law 25
APPLICABLE BY DEPLOYMENTQUÉBEC · PRIVATE-SECTOR PRIVACY
Map privacy governance, impact assessments, consent, transparency, incident response, retention, destruction, and cross-border processing for Québec operations.
PIPEDA
APPLICABLE BY DEPLOYMENTCANADA · FEDERAL PRIVATE-SECTOR PRIVACY
Implement accountability, purpose limitation, consent, access, retention, and safeguards appropriate to the sensitivity of personal information.
ISO/IEC 27001:2022
PRIMARY CERTIFICATION TARGETINFORMATION SECURITY MANAGEMENT
Use a risk-based information security management system as the certifiable foundation for people, process, technology, suppliers, and continual improvement.
ISO 28000 / 28001
DOMAIN ALIGNMENTSUPPLY-CHAIN SECURITY
ISO 28000:2022 supplies security-management requirements; ISO 28001:2007 adds supply-chain security assessments and plans. Both complement rather than replace ISO/IEC 27001.
SHARED CONTROL BASELINE
- 01Data residency and deployment choice
- 02Encryption and managed secrets
- 03Least privilege and tenant isolation
- 04Auditable agent and human decisions
- 05Retention, deletion, and export controls
- 06Human approval and bounded model scope
07 — Proof
Evidence across physical operations.
Not a lab claim: live devices, customer-owned workflows, and operators using the system where temperature, traceability, and uptime matter.
Production snapshot · 2 September 2026
OBSERVED OPERATING SCALE · NOT PROJECTIONS
- 275
- Live devices
- 3
- Countries with live data
- 16+
- Fermenters in one production estate
- 4
- Operating contexts represented below
Operating evidence 01 · Process
“Two years in, and AEDDIX systems have become integral to our success, helping ensure the quality, consistency, and reliability of our operations. Over 16 fermenters and our taproom, the platform has proven simple to use and highly effective in supporting our team day to day.”
“From infrastructure migrations and custom application development to cold chain monitoring and operational digitization, the team has consistently supported our vision every step of the way.”
“With telematic GPS fleet monitoring and custom cold chain management devices, we now benefit from stronger traceability, better operational control, and greater confidence in our logistics environment.”
“Over the years, the platform has delivered real value through stronger customer engagement, better data transparency, and reliable operational support.”
Let the estate ask for a human only when it matters.
Bring one messy fleet and one costly exception. We will show where context is added, why policy allows an intervention, and how the action protects the operation.