Prototype v0.9 · InsighTwin™ module page · T1 template instance #8
Layer 4 · Operational Intelligence · InsighTwin™

Why did this happen? In one sentence.

InsighTwin™ explains your plant's numbers in language a Plant Head can use, not just a data scientist. Why OEE dropped 2 points on Tuesday. Why Line 3 scrap spiked at 14:00. Why A-shift is consistently behind B-shift on the same product. Cause, effect, and a recommended action — in plain English.

Plain-English explanations
Auto cause-effect
Plant Head dashboard
// INSIGHTS · TODAY
23 surfaced
// ACTIONABLE
8 of 23
// MD BRIEF · 09:00
delivered
23/day
Insights surfaced per plant
8/23
Median actionable insights
Plain-English
Cause-effect explanations
Plant-floor
No data science required
In 30 seconds

What InsighTwin™ does.

// 01

Auto-generates the variance report.

Every week, every shift, every line. Why was today different from yesterday? InsighTwin writes the answer.

// 02

Explains in plain English.

Not charts. Not p-values. "OEE dropped 2 points because Line 3 changeover ran 14 minutes long because M-302 bearing temp triggered a pause."

// 03

Drills from MD-grade to operator-grade.

Same insight, four levels of detail. MD sees the headline. Plant Head sees the cause. Maintenance sees the bearing log.

Built for

Plant Heads who got tired of data scientists telling them what their plant did. Not why it did it.

Most Indian plants have a dashboard somewhere with green and red numbers. Almost none have a dashboard that explains why those numbers are green or red. InsighTwin closes that gap. The Plant Head opens it at 08:30 and gets the night-shift story before walking the floor.

Primary buyer

Plant Head · COO · MD · Quality Head

// THE PAIN

We have dashboards. We have KPIs. We have weekly review meetings full of charts. What we don't have is a clear answer to "why did this happen?" Without that, every week is the same conversation: red number, blame meeting, no resolution. We need an analytics layer that explains itself.

// SUCCESS METRICS

23
Insights/day
8
Actionable
0
Charts to interpret
How it works

Four steps.

// 01 · OBSERVE

Reads everything below it.

SmarTwin live data, VolTwin energy, PredicTwin alerts, AssetTwin asset state. One unified plant view.

// 02 · CORRELATE

Finds the why.

Cross-source pattern recognition. A 2-point OEE drop correlates to a 14-minute changeover anomaly. The bearing alert from PredicTwin caused the changeover.

// 03 · EXPLAIN

In plain English.

Auto-written variance commentary. No charts. No data science. Plant-floor specific language.

// 04 · DELIVER

To the right person.

MD scorecard at 09:00. Plant Head dashboard at 06:00. Shift supervisor mobile at end-of-shift.

Capabilities

What InsighTwin™ ships with on day one.

AE

Auto-Generated Explanations

Daily variance commentary auto-written. Plain English. Plant-floor specific.

CE

Cause-Effect Detection

Cross-source correlation. Bearing alert → changeover delay → OEE drop. Linked automatically.

DR

Drill-Down Levels

MD-grade headline → Plant Head cause → operator-grade detail. Same insight, four resolutions.

BR

Daily Briefings

MD scorecard 09:00 IST. Plant Head dashboard 06:00. Auto-delivered to inbox + dashboard.

SH

Shift Comparison

A-shift vs B-shift on same line, same product. Surfaces consistent gaps in workmanship.

AN

Anomaly Surfacing

What's unusual today that wasn't yesterday. Surfaced before it becomes a problem.

RC

Root Cause Suggestions

Not just "the OEE dropped." Why the OEE dropped, plus the suggested action.

NL

Natural Language Q&A

Ask "why is Line 3 scrap up this week?" in the dashboard. Get a written answer in seconds.

BE

Cross-Plant Benchmark

Why does Plant A make this product better than Plant B? InsighTwin says.

RR

Recurring Issue Detection

The same root cause showing up across multiple weeks. Pattern surfaced.

AC

Action Tracking

Suggested action → assigned → status. Closes the loop from insight to outcome.

EX

Export to Review Meeting

PowerPoint or PDF export of the week's top 5 insights for the weekly review.

Plant-floor truth

Numbers we can defend.

23/day
Insights auto-surfaced per active plant deployment.
8/23
Median actionable — Plant Heads act on within 24 hours.
0
Charts to interpret. Every insight is written, not drawn.
↓ 34%
Time spent in weekly review meetings, post InsighTwin rollout.

Source: Wistwin internal benchmark across active deployments, January 2024 – April 2026.

Pairs with

InsighTwin™ in the stack.

FAQ

Questions buyers actually ask.

InsighTwin™ is the analytics layer of the Wistwin® digital twin platform that explains your plant's numbers in plain English. It reads from every other Wistwin module (SmarTwin telemetry, VolTwin energy, PredicTwin alerts, AssetTwin asset state) and auto-generates daily variance commentary that any Plant Head can use immediately. No charts to interpret. No p-values. Just "this happened because of this — and you should do this."
Both. InsighTwin uses a hybrid: rules-based pattern matching for known cause-effect chains (bearing alert → changeover delay → OEE drop), and LLM for natural-language explanation generation. Every LLM output is grounded against actual sensor data and verifiable. We do not hallucinate plant numbers.
Plant Head dashboard updates at 06:00 IST with night-shift variance. MD scorecard delivered to inbox at 09:00 IST with the day's top 5 actionable insights. Shift supervisor mobile updates at end-of-shift handover. Customizable per customer schedule.
Yes. The dashboard includes a natural-language Q&A field. "Why is Line 3 scrap up this week?" "Why is A-shift behind B-shift on Product X?" "What changed in energy consumption Monday vs Tuesday?" InsighTwin answers in seconds, grounded against actual data.
Cause-effect attribution accuracy varies by complexity. Simple chains (alert → downtime → OEE) hit 90%+. Multi-factor variances (where 3 separate causes contribute) sit around 70-75% on first attempt — InsighTwin flags multi-factor cases explicitly and asks the Plant Head to confirm. We always show the underlying data so the Plant Head can verify.
Pilot: 6-8 weeks. Plant rollout: 10-12 weeks. The model training period is 8 weeks of live data before InsighTwin's plant-specific insights are at full quality. Generic Indian-plant patterns work from week 1; plant-specific patterns sharpen over 8 weeks.
Yes. AWS Mumbai by default. On-prem available. The LLM inference component runs on India-resident Anthropic or in-house models for customers requiring strict residency. DPDP-compliant. ISO 27001 in progress.
InsighTwin is priced as a Layer 4 module on the Wistwin platform — Optimize tier. Pricing is per-plant per-year. Indian customers typically see payback in 8-10 months through reduced weekly-review meeting hours and faster issue resolution.
AssetTwin + at least one telemetry source (SmarTwin minimum). Adding VolTwin and PredicTwin dramatically expands what InsighTwin can correlate across. The more modules live, the richer the insights.
Yes. One-click PowerPoint or PDF export of the week's top 5 insights, with cause-effect chain, recommended action, and source data references. Used in production by Plant Heads across pharma, auto, and chemicals deployments.
The 4-minute benchmark

Where does your plant stand on plant analytics?

Take the DMM Check™. Get a personalised readiness score, a benchmark against 200+ Indian plants in your sector, and the top three priorities InsighTwin™ should activate first.

India work emails only. No sales call until you ask.