The same problemsacross different industries
Waste collection, equipment rental, warehouse handling and hazardous work involve different operations, yet many of the underlying problems recur.
AI moves from conversation to action
Large language models allow software to interpret goals, plan steps and use tools, enabling AI to carry out sequences of work. In digital environments, it works with documents, databases and interfaces: states can be queried and many actions can be reversed.
Physical operations are different. They involve people, equipment, vehicles, materials and places. Their states change continuously and only enter the system when sensed. Actions have real consequences: cutting power, restricting access or dispatching someone to a site. Before AI can act, it must reliably establish what is happening, where it is happening and who or what is involved. It can then determine the action, assign responsibility and verify completion.
Three gaps between AI and the site
Lack of visibility
- On-site reality
- People and mobile assets lack continuous location records. Equipment and environmental conditions rely on manual readings or occasional inspections. Short sensor battery life makes sustained coverage difficult.
- Effect on AI
- Objects it cannot see are missing from its assessment
Unreliable context
- On-site reality
- Data sits in separate systems with inconsistent object identifiers and timestamps
- Effect on AI
- Assessments lack evidence, conclusions cannot be checked and responsibilities are unclear
Limited control
- On-site reality
- Instructions travel by phone, group chat or paper work order, with results reported manually. Permissions and safety prerequisites remain in policy documents rather than executable checks.
- Effect on AI
- AI cannot confirm whether an instruction arrived, was carried out or achieved the intended result
In dashboard-centred systems, people fill these gaps: they judge data using experience, phone the site to confirm conditions and compile records afterwards. AI needs those capabilities in its interfaces, with complete, traceable information and defined ways to act.
Five recurring operational problems
Waiting
- Typical situations
- Waste-transfer trucks queue at stations, freight vehicles wait to load or unload, and production lines wait for materials or maintenance staff
- Missing spatial facts
- Vehicles, goods, loading bays and work schedules do not share a common timeline
Utilisation
- Typical situations
- Idle equipment and local shortages coexist in construction and healthcare, while productive working time is unclear
- Missing spatial facts
- Whether people and equipment are present, in use or available
Authorisation
- Typical situations
- Who may enter a restricted area, and who is authorised for high-voltage or hot work
- Missing spatial facts
- A real-time link between identity, qualifications and location
Accountability
- Typical situations
- Who was present or absent when an event occurred, and which party caused a delay
- Missing spatial facts
- Continuous movement and event records that can be replayed
Settlement
- Typical situations
- Outsourced services are paid by hours, trips or coverage, and equipment by productive operating hours
- Missing spatial facts
- Measurement methods and supporting evidence accepted by both parties
Records support work coordination and reconciliation, not punitive performance scoring or fines. Explain what is recorded and how it will be used before deployment.
Why the nature of the problem has changed
These gaps were often treated as efficiency problems, addressed through management and manual coordination. As AI with decision-making and execution capabilities enters physical operations, the requirement changes. AI cannot fill missing data with experience and phone calls; it needs infrastructure designed to make the site accessible to it.
Adapted from chapter 1 of the IoT4Agent white paper (Lindos, v1.1, October 2026).Read the IoT4Agent white paper →
Related solutions
Transport coordination
Split or combine orders and plan vehicles and routes.
- ↓Empty-running distance
- ↑On-time deliveries
Equipment rental operations
Match equipment with operators and reconcile actual working time.
- ↓Equipment idle time
- ↓Billing disputes
Subcontractor performance
Check working time against location records and flag entry into hazardous areas.
- ↓Pay reconciliation disputes
- ↓Safety response time
Measure outcomes consistently: agree the indicators and record a baseline before starting, then compare results using the same definitions.
Related articles
Start with one workflow
Choose a frequent task with a clear operational problem and measurable results, then progress through five practical steps.
From dashboards to AI-ready operations
IoT has connected equipment and brought data to dashboards. Its next role is to make physical operations readable, trustworthy and actionable for AI.
Start with one site and one workflow
contact@lindos.cn
Deploy into your existing site without rewiring, alterations or interrupting operations. Tell us which workflow you want to improve first. Together, we will agree the available data, initial scope and measures of success.
Talk to us
Tell us about your site and the workflow you want to improve. We will arrange a conversation to agree how to get started.
contact@lindos.cn →02Request information
Explore solutions for your industry, the sensors involved and a practical starting scope.
contact@lindos.cn →03Partner with us
Develop and deliver a joint solution with clearly agreed roles and responsibilities.
contact@lindos.cn →