TSAI — Time Series AI Platform

What TSAI Is

The Data Infrastructure
AI Actually Needs.

Every AI project needs data. But not just data — the right data, organized correctly, with the right context attached. Most industrial operations have thousands of sensors and years of history. Almost none of it is structured in a way that AI can reason about effectively. TSAI solves that problem.

TSAI connects to your existing automation systems and begins capturing operational data in a time-series format purpose-built for AI consumption. But data capture alone isn't enough — the platform is specifically designed to capture the human context that machines can't record on their own.

When an anomaly occurs, TSAI detects it. But it also gives operators the ability to annotate what they observed — what changed, why it happened, what they did about it. That human layer is what transforms raw sensor data into something AI can actually learn from.

The result: a living operational record that combines machine precision with human knowledge — structured, contextualized, and ready to feed today's AI tools.

The Context Problem

Today's AI is powerful — but it needs context to be accurate. A model that sees a magnetite spike without knowing that Mine A coal was running that shift will draw the wrong conclusion. TSAI captures both the event and the explanation, so the AI has the full picture.

Purpose-Built for Mining & Mineral Processing

TSAI is not a generic IoT platform adapted for industry. It was designed from the ground up around the operational patterns of preparation plants, longwall operations, and mineral processing circuits — the event structures, the shift rhythms, the tag hierarchies, the delay classifications that matter in this industry.

How It Comes Together

1 — Organize

Coal Analytics Portal — organized analysis modules

Analyses organized into targeted modules — each with its own data, methodology, and outputs.

2 — Add Context

Knowledge model — process and operational context

The knowledge model captures plant structure, process rules, and operational context that AI needs.

3 — Investigate

Explore the full operational picture — timelines, sensor data, events, and shift context on one screen.

4 — Focus

Heavy media analysis — focused analytical deep-dive

Targeted analysis on specific processes — heavy media, blend, operator behavior — where the money is.

5 — Operationalize

Blend operationalization — actionable recommendations

Findings become action plans — specific, trackable steps that change behavior on the plant floor.

Data Capture

Connect to Anything.
Capture Everything.

TSAI connects to your existing automation infrastructure without requiring changes to your control systems. If your plant produces it, TSAI can capture it.

Machine Data

Direct connection to PLCs, DCS, SCADA systems, and process historians via standard industrial protocols.

Operator Context

The human layer that makes AI work. Operators annotate events, add observations, and capture the operational knowledge that sensors can't record.

Integrated Sources

TSAI joins time-series sensor data with operational records from your existing systems — creating the integrated dataset that no single source can provide alone.

On-Premise. Your Network. Your Data.

TSAI runs on-premise at your site. No cloud dependency. No data leaving your facility. Works on your plant network with your existing security infrastructure. IT approval is straightforward because there's nothing external to approve.

Mine

Convey

Process

Analyze

Heavy media operations timeline analysis

AI Readiness

Data Isn't Enough.
Context Is Everything.

Today's AI tools are capable of remarkable things — but only when the data fed into them is correctly structured and richly annotated. TSAI is designed specifically to create that foundation.

Event Detection

TSAI continuously monitors your process and automatically detects anomalies, state changes, and operational events — SG drift, magnetite spikes, chute plugs, blend transitions, equipment stops. Every event is timestamped and tagged.

This gives AI models clean, labeled events to reason about — not raw time-series noise.

Operator Annotation

When an event fires, the operator can attach context — what coal source was running, what changed upstream, what they observed on the floor. Voice input or text. Simple enough to use mid-shift.

Without this layer, AI sees a magnetite anomaly. With it, AI understands "Mine A coal was transitioning — this always happens during that blend."

Structured History

Over time, TSAI builds a structured operational history: annotated events, shift context, delay classifications, operator observations — all linked to the machine data that triggered them. This is the training dataset that makes your AI models specific to your operation.

Generic AI models trained on industry data can't match a model trained on your plant, your operators, your coal.

LLM Integration

Because TSAI's data is correctly structured and contextualized, it feeds directly into large language models to generate natural language shift reports, anomaly narrations, and operator coaching — automatically, every shift.

The AI doesn't just find the pattern. It explains it in plain language to the person who needs to act on it.

Analytical Toolkits

Targeted Solutions.
Built on Your Data.

Once TSAI has your data correctly captured and contextualized, we build targeted analytical toolkits on top of that foundation. These aren't generic dashboards — they're operationally specific analyses designed to answer the questions that actually cost you money.

Heavy Media & Magnetite Effectiveness Magnetite is expensive. In one deployment, over a quarter of all magnetite additions were unnecessary or counterproductive — operator-driven rather than process-driven. TSAI identifies exactly when and why this happens: which operators, which shifts, which blend conditions trigger over-addition. K-means clustering reveals distinct operator behavioral profiles that are invisible to conventional metrics. Shift-change behavior — where handover patterns drive unnecessary additions — is quantified and addressable.

SG drift classification and control limit monitoring
Unnecessary addition detection and quantification
Per-operator profiling and behavioral clustering
Shift-change pattern analysis

Blend Optimization When multiple coal sources feed the same preparation plant, blend composition drives yield and quality outcomes in ways that are rarely quantified. TSAI tracks mine feed contributions in real time and correlates blend composition against yield, ash, sulfur, and BTU — building models that identify the optimal blend for any given quality specification. The result is a data-backed recommendation for maximizing throughput within your compliance constraints.

Real-time blend composition tracking by mine source
Yield and quality modeling by blend regime
Constrained optimization for ash and sulfur compliance
Blend-to-quality correlation and prediction

Operationalization Checklist & Scorecard Finding the problem is only half the work. For each analytical finding, TSAI generates a targeted operational plan — a dynamic checklist of what needs to change, shift by shift, operator by operator. That checklist becomes the scorecard. Compliance is tracked, not just recommended. This is where insight becomes accountability, and where the findings from the other toolkits stop being reports and start changing behavior on the plant floor.

Dynamically generated operational plans from findings
Per-shift, per-operator checklists tied to specific issues
Scorecard tracking of checklist compliance
Direct feed into Performa for shift accountability

How You See It

Intelligence Delivered
Where Your Team Works.

TSAI surfaces its analytics through the interfaces your team already uses — no new software to learn, no browser required for most functions, no IT approval needed for daily use.

Industrial Displays

Screen rotation systems cycle through HMI emulations, performance dashboards, and KPI boards on plant floor and control room displays — 24/7, no interaction required.

Excel Add-In

A purpose-built Excel add-in gives operational staff and supervisors direct access to TSAI data from within the tool they already use every day. Bulk data operations, scorecard access, and reporting without leaving the spreadsheet.

PowerBI & Web Dashboards

Executive and management reporting through PowerBI. Interactive operational dashboards through the web interface. Both draw from the same TSAI data layer — one source of truth for every level of the organization.