Strengthening Xempla as the system of decisions for operations and maintenance
‍

This month at Xempla focused on strengthening how operational decisions are made across assets and facilities.

Facilities operations generate large amounts of information — maintenance history, inspections, work orders, documentation, and operational signals. But these signals often remain fragmented across workflows, making it difficult for teams to understand the real state of assets and act confidently.

February’s updates move Xempla further toward its core role as a system of decisions for operations and maintenance teams — connecting operational context, AI agents, and human supervision so that decisions are made with greater clarity and confidence.

Two developments were central to this progress: the introduction of the Model Context Protocol and the launch of Assurance Scores for assets and locations.

Model Context Protocol
‍

Shared context for agents and human decision-making
‍

Operational decisions depend heavily on context — design documentation, maintenance history, operational conditions, and past investigations.

Until now, much of this context has been fragmented across different workflows and tools.

In February, we introduced the first version of Xempla’s Model Context Protocol (MCP). This capability allows AI agents and human supervisors to operate using the same contextual understanding of an asset.

It enables:

• Agent-to-agent collaboration when resolving operational issues
• Shared asset context across decision workflows
• Human supervision with full visibility into how decisions are made

The protocol also powers Luma, Xempla’s on-site technical assistant. Technicians can ask questions in natural language and receive responses grounded in the asset’s operational context.
‍

Why it matters

For facilities teams, better decisions come from better context. By structuring asset knowledge and making it accessible across workflows, the Model Context Protocol helps ensure that both humans and AI agents operate from the same source of truth.

This allows Xempla to move beyond simply tracking operations toward supporting decisions across the lifecycle of assets.

Assurance Scores for Assets and Locations
‍

A single signal for operational confidence
‍

Operations teams often evaluate performance through multiple disconnected metrics — preventive maintenance compliance, ticket closures, recurring issues, inspections, and documentation quality.

While each metric provides insight, they rarely offer a clear answer to a fundamental question: How confident can we be in the performance of this asset or facility?

To address this, Xempla introduced Assurance Scores for both assets and locations.

For assets, the score considers operational factors such as preventive maintenance adherence, quality of triaging and maintenance execution, documentation completeness, and closure quality of reactive work orders.

For locations, the system aggregates signals like ticket resolution performance, operational risks, and compliance activities across the facility.
‍

Why it matters

Assurance Scores consolidate multiple operational activities into a single indicator of confidence, similar to how reliability engineers evaluate system performance holistically. For facilities and O&M leaders, this provides a clearer understanding of where operational attention is needed.
‍

Outcome

Teams can quickly identify:

• Assets with declining reliability signals
• Locations where operational processes need improvement
• Areas where maintenance practices are strengthening overall performance

This allows Xempla to surface the signals that matter most for operational decision-making.

Extending Decision Intelligence into Existing Platforms

February also marked Xempla’s first integration with an external operational platform.

Through this integration, Luma’s intelligence layer can be embedded directly within third-party systems, allowing users to access asset insights and operational context without leaving their existing workflows.

This expands the reach of Xempla’s decision system while fitting naturally into how teams already operate.

🔮 What’s Next — March Focus

Looking ahead, March development focuses on strengthening governance and operational data quality within the decision system.
‍

Governance Layer for Agent–Human Collaboration

As AI agents become more active in operational workflows, clear governance becomes essential.

We are introducing a governance layer for agent-to-human handovers, ensuring that decisions escalated to humans carry full context and reasoning.

At the core of this work is the development of an asset context graph, which captures operational knowledge and decision history across both human and AI actions.

This ensures that decision-making remains transparent, auditable, and structured.
‍

Stabilizing the System of Record

To support a strong decision system, operational data must remain reliable and easy for technicians to capture.

Current work focuses on:

• Improving technician workflows and UI efficiency
• Reducing the time required to record operational data
• Embedding automated data quality checks at the source

These improvements help ensure that the operational signals feeding Xempla’s decision system remain accurate and trustworthy.
‍

R&D: Image Intelligence for Maintenance Workflows

We are also exploring how images captured during maintenance activities — inspections, ticket creation, and work order closure — can provide richer operational context.

This research focuses on using visual evidence to strengthen auditability, context, and decision support within maintenance workflows.

🌍 The Bigger Picture

February’s developments reinforce Xempla’s evolution as a system of decisions for facilities and operations teams.

The Model Context Protocol connects operational knowledge across workflows. Assurance Scores bring clarity to asset and facility performance. And upcoming governance capabilities ensure that AI-driven workflows remain transparent and accountable.

Together, these capabilities help organizations operate with better context, clearer signals, and stronger operational decisions.

‍