Data as the Nervous System of the OSH Management System in Large Enterprises
➡ From isolated dashboards to actionable decisions: this is how OSH scales with Data-as-a-Service, embedded BI, KPI governance and Edge ingestion (Sabentis– BOSCH ).
The organizations that will lead the future understand that occupational health and safety (OHS) is not just about compliance: it’s about architecture . These companies are designing systems where data flows like a nervous system , connecting metrics, context, and action in real time to manage risk with precision. In countries like Colombia, this translates into concrete requirements: we see it in OHS regulations and the OHS Management System (SG-OSH ).
Along this path, it’s common for each department to have its own dashboard. The difference lies in what’s behind it: the most advanced companies are consolidating a single data map and an auditable KPI → Evidence → Action → Closure chain , where the indicator isn’t just observed: it’s managed .
⚡Executive Summary (30 seconds)
What we at Sabentis work on together with companies that already work on occupational safety and health at scale can be summarized as follows:
1. That all indicators speak the same language
KPIs are no longer an isolated formula; they now have a clear, shared, and stable definition across the entire organization. It’s important to be clear about which key performance indicators (KPIs) you should be measuring in occupational health and safety .
2. Turn evidence into useful information
Documents are no longer simply filed away; they are structured so that they can be traced, related, and used in decision-making.
3. Bring analytics into everyday life
The analysis lives within the work environment, integrated and with the appropriate permissions for each role.
4. Operate the data as a continuous service
KPIs are not published and forgotten; they are constantly managed, updated, and validated.
5. Integrate industrial data into the same model
Information received from sensors or IoT environments adds value when it enters the same governance system and does not function as a separate platform.
The result: a clear and operational cycle where the signal is captured, structured, analyzed, acted upon, and evidence is recorded.
How are the most advanced organizations using data in occupational risk prevention?
Interviewer: Payam, when you work with large organizations to structure their data strategy in OSH, what is the first thing you align?
Dr. Payam (Head of Data & AI, Sabentis): The first step is understanding that the OSH management system is not a document repository, but a system that must function in real time. And we built that on three fundamental pillars.
1. Corporate Dictionary of KPIs
When an organization has dozens of centers, a “KPI” ceases to be just a number: it becomes a decision interface . More mature companies are creating a corporate dictionary where each KPI has:
- A formal definition
- Explicit business rules
- A time horizon, filters and regulatory context
- owner, lifecycle and version control
2. Evidence as data (structured, contextual and auditable)
The second step is to transform the evidence into computable evidence .
Center → Place → Position → Exposure/Risk → Event → Measure → Responsible person → Evidence → Status
This enables the highest value-generating operating standard: drill-to-proof .
3. Analytics within the workflow (embedded BI + security by design)
The third movement is distribution: bringing analytics to the workplace. When BI resides externally, it’s consulted; when it resides internally, it’s used for operations .
That’s why the pattern we’re seeing in large corporations is clear:
- Power BI Embedded integrated into Sabentis
- RBAC + RLS to segment by role, center, country, holding
- access auditing as a corporate standard
From “Reporting” to Data-as-a-Service: the indicator as a product, not as a report
Interviewer: What exactly does DaaS applied to OSH mean?
Dr. Payam: DaaS is operating OSH as a continuous service that provides guarantees. That means:
- Stable metrics (defined, documented, versioned)
- Easy access (embedded in the same flow)
- End-to-end traceability (the KPI is aggregated back to the primary evidence)
In OSH, simply seeing the number is not the goal. The goal is for the number to enable actionable decisions.
Which site? Which workstation? Which exposure? Which event explains it? Which action is open? Who validates and when does it close?
Minimum architecture for OSH
At Sabentis, we express this minimal architecture as a layered architecture:
1. Model layer: organizational context and inheritance
In large companies, context is not a detail: it is the driving force.
- The organizational structure is the map
- The workstation is the link between the physical and the operational
- Inheritance prevents duplicate assessments and fragmented exposure data
When this layer is well designed, the data ceases to be a “record” and becomes a nervous system .
2. Semantic layer: BI and AI speaking the same language
Companies building the future are solving a key problem: ensuring that BI and AI models use the same semantics.
That requires:
- datasets with stable definitions
- dimensional model / semantic layer
- metrics governed as “single source of truth”
3. Access layer: real security and mass distribution
At scale, analytics only works with corporate security:
- RBAC : permissions by role
- RLS : Segmentation by center/country/holding
- traceability of who accesses what and when
This is what makes embedded BI an enterprise capability.
KPI Factory: industrializing the indicator as if it were software
Interviewer: How do you prevent the KPI from becoming fragmented over time?
Dr. Payam: Treating it as a product with a life cycle . At Sabentis we call it KPI Factory: an industrial way of producing indicators with change control.
A mature KPI Factory includes:
- global catalog of KPIs
- deliveries by package (by version)
- definition template (question, formula, rules, filters by country)
- Functional validation + technical validation + QA before publishing
The key for multinationals is to accept a reality: there is no single “rate” ; there are different versions of the rate in different contexts. Industrializing this prevents drift and ensures comparability.
DataOps in OSH: where strategy becomes operation
The data strategy in OSH is not demonstrated in a diagram; it is demonstrated in operation.
The companies that are building the future are professionalizing:
- managed analytical environments
- standardized validation and publication processes per tenant
- quality criteria (data quality and product quality KPIs)
- continuous governance (incidents, tickets, internal SLAs)
When DataOps enters OSH, data ceases to be a “project” and becomes a sustained capability .
AI in OSH: When Artificial Intelligence Turns into Action
Interviewer: What role does AI play in this architecture?
Dr. Payam: AI provides an advantage when it amplifies three things:
- reduces friction (capture/structuring)
- prioritize with context (not just historical)
- It explains why and is integrated into decisions.
- lands in workflow (action + evidence)
In practice, a useful AI doesn’t end with a simple recommendation: it ends in a closed loop.
The next leap: EdgeOne + BOSCH as a new industrial signal layer
Interviewer: Where does Edge/IoT fit into the data strategy?
Dr. Payam: It fits in as another source within the nervous system, always under the same governance and model. That’s why EdgeOne is designed as governed intake : capturing in the field, normalizing, correlating, and activating action.
At Sabentis, EdgeOne appears as a new data category linked to the agreement with Bosch for capturing and transmitting data from work environments. The idea is simple and powerful: converting industrial signals into governed safety and health (OSH) events , and from there into decision-making and evidence.
“If the alert leads to action and evidence, prevention becomes operational at scale.”
Three tests of maturity: when a KPI truly governs
Payam summarizes it with three simple (and very demanding) tests:
- Reproducibility : The KPI is consistent across teams and centers
- Drill-to-proof : from graph to evidence + responsible person + status
- Actionability : the KPI activates flow; it does not just inform
When these conditions are met, OSH stops being “measured” and starts to be implemented .
The turning point: when data governs occupational safety and health
The question that defines the standard to which the organizations that will lead the future are evolving is:
Can we answer in seconds, with evidence: what critical risks are open today, who is responsible and how the closure is progressing, with complete traceability of KPI → action?
When the answer is “yes”, the data ceases to be support and becomes strategic infrastructure : the nervous system of OSH.
