Welcome to Seeq Summit Mumbai 2026

2

09:00 – 10:00 – Registration and Networking

10:00 – 10:10 - Welcome & Kick Off

10:10 – 10:40 - Opening Keynote: Human Expertise, Amplified - Dustin Johnson, CTO at Seeq

10:40 – 11:25 - Leveraging Seeq to Improve Asset Reliability - Shell India Markets Pvt. Ltd.

11:25 – 11:45 - From Data to Decisions: Right First Time and Higher Efficiency with Seeq in Pharma - Seeq

11:45 – 12:15 - Networking Break

12:15 – 12:45 - Seeq in Production of Biopharmaceuticals - Lonza

12:45 – 13:15 - Trusting operational data: the true prerequisite for reliable Industrial Advanced Analytics - Ematica

13:15 – 14:15 - Lunch & Networking Break

13:15– 14:15 - Live Demo: Stop by the Seeq Genius Bar to see live demos and ask questions

14:15 – 16:00 - Hands-On AI Workshop

16:00 – 16:30 - Optimizing Boiler and Steam Performance Using Seeq - Indorama Ventures

16:30 – 17:00 - From Machine Learning to Operational Decisions: Bringing Predictions to Where Decisions Happen - Seeq

17:00 – 18:00 - Networking & Refreshments

Agenda

4

SEEQ SUMMIT SEEQERS

5

Points to Note During the Day

▪ Safety

▪ Wi-Fi

o Network Name: SEEQ

o No password required

▪ Breaks and Lunch

o Will take place in the foyer

▪ Networking Reception

o Will take place in the EL rooftop bar

6

Ask Your Questions Using Slido

Scan the QR Code to join our active Slido Session where you can ask questions throughout questions throughout presentations. The presenter will answer them at the end of their session.

Questions can be submitted anonymously or with your name attached.

8

Please Welcome our Next Speaker

DUSTIN JOHNSON

CTO, Seeq

Human Expertise, Amplified

Dustin Johnson CTO & FOUNDING PARTNER

10

The pressure on operations has never been greater.

AI is positioned as the solution, but is it?

Speed and decision quality matter more than ever.

11

“We’re drowning in data, but starved for information.”

12 Data

Information

Contextualize

Insights

Discover

Decisions

Brainstorm

13

Built By Industry. Focused On Experts.

Expertise Amplified by AI.

14

15

Enterprise Monitoring

Scale Across AssetsSelf-Service Analytics Operationalize

Condition Monitoring

TriageInvestigate

Continuous Improvement

16

Decision Capital

Decision Capital

The accumulated capability of an organization to make good decisions

Identify & Prioritize

Opportunities

Connect & Codify

Capture & Quantify

Outcome

Investigate,

Decide & Act

Information

Context

Lessons learned

Decision history

Institutional knowledge

Expertise

17

AI Activates Decision Capital

AI helps synthesize context

AI helps surface recommendations

AI helps connect information

AI helps find information

Identify & Prioritize Opportunities

Connect & Codify

Experts at the

Center

Capture & Quantify Outcome

Investigate, Decide & Act

18

Identify & Prioritize Opportunities

Connect & Codify

Capture & Quantify Outcome

Investigate, Decide & Act

19

20

Corporate Agentic

Ecosystem

Process Data

SAP Maximo

SharePoint

Databricks

Email

Oracle

ServiceNow

Human Expertise Is The Advantage. Amplify it.

Dustin Johnson CTO & FOUNDING PARTNER

22

Any Questions?

Please raise your hand and a team member will bring you a microphone.

OR

Scan the QR code to submit your questions using Slido.

23

Please Welcome our Next Speaker

PAOLO BRAIUCA

Industry Leader, Pharma and Life Sciences, Seeq

From Data to Decision: Right First Time and Higher Efficiency with Seeq in Pharma

SEEQ IN PHARMA

Paolo Braiuca, Industry Leader Pharmaceuticals

26

The Problem: Raw Data -> Quality Decisions “We keep having downstream product quality issues and can’t connect the dots”

“Can you review all product data & our monitoring program and sort this out?

We can’t have any more investigations & need answers ASAP!”

Where do I start?

Lab Data

MES

The data exists, but the product & batch risk context is

fragmented through the lifecycle

28

Seeq for Pharma supporting compliant decisions

Seeq helps teams move faster without moving the source of truth.

1 Preserve the validated source

• Leave data where it is — no duplication, no ETL

• Connect context across process, batch, lab, quality, and asset data

• The system of record stays in place

2 Make the reasoning traceable

• Auditability and analysis versioning

• Documented methods

• Reproducible evidence

3 Scale governed workflows

• Controlled release of analyses

• Consistent templates and role- appropriate views

• Reusable site patterns

Faster decision-making, with evidence that can be reviewed, explained, and defended.

29

What decisions?

New Product Introduction

• Process understanding and

optimization

• Process “fingerprinting”

• Tech transfer

Fewer engineering Batches

Faster time to market

Operational Control and Efficiency

• Real-time process monitoring

• Real-time operational efficiency

monitoring

Fewer quality deviations

Increased throughput

Capacity utilization

Cost control

“Automated” Paperless QA

• Deviation investigation and

documentation

• Batch reporting and release by

exception

• Continuous process verification

(CPV/OPV)

• Asset qualification

Faster batch release

Audit readiness

Cross-site consistency

Reduced quality risk

Equipment and Facilities Management

• Asset health

• Usage-based and/or predictive

maintenance

• Cleanroom / sterile environments

monitoring

• Utilities and Wastes

management

Increased equipment availability

Higher production capacity

30

From Blind Spots to Actionable Insight

The Challenge

Something is wrong:

How to see it clearly?

Why is it happening?

What to do about it?

Most tools answer the what

- OEE metrics,

- Alarms,

- Batch summaries

but rarely explain the why.

5–15% of recoverable efficiency (often more!) exists in most

pharma manufacturing operations

Holistic Visibility

A unified view across equipment, batch phases, operations, scheduling, quality; not

siloed metrics.

Root Cause Investigation

Go from 'what happened' to 'why it happened': fast, without manual data wrangling

across systems.

Targeted Improvement

Apply changes with confidence and monitor whether they're working, closing the

improvement loop.

Real-Time Drift Detection

Catch deviations early before they become losses — alarm intelligence, not alarm

fatigue.

31

One GMP decision depends on more than one data stream

T I M E

DS Batch DP Batch FP Release LT Stability

PI / Process Real Time time-series

MES Context batch ID, product ID, run status, stage

LIMS / Release Lab results arrive later: assay, dissolution

QMS Events events may be real-time or delayed

LIMS / Stability critical stability pulls arrive months later

Unified Analysis Story

• Aligned in time

• Connected by relationships

• Contextualised for deeper insight

• One analysis, complete context

32

First, make the batch understandable

O N E B A T C H , A S S E M B L E D C O N T E X T

Batches and phases CPPs

CQAs Qualified ranges

Equipment History

Context turns a trend into meaningful information

33

34

• Always-on monitoring on the signals that matter.

• Monitor leading process indicators alongside final CQAs.

• Combine specifications, qualified control limits, and run rules.

• Create contextualized events instead of isolated alarms.

Learn from events: INVESTIGATE and improve your control strategy

Step 2 – Define your Control Strategy Detect the exception when there is still time to act

35

36

Step 3 – Scale and Monitor by exception

• Extend the control strategy logic across the product portfolio and manufacturing assets

• Monitor in real time at scale

• Manage by exception

• Evaluate risks opportunities and prioritization

• Automate reporting

Move from one-off, retrospective and manual to ALWAYS-ON BY EXCEPTION

Focus on what matter and act as it happens

37

38

Better process understanding

Fewer quality deviations

Higher yield/throughput

Reduced investigation and reporting time

Increased equipment availability

Operational cost reduction

Audit-readiness

This is now

Faster decisions drive business value

30% reduction of engineering batches

75% reduction of quality events in early commercial batches

22% yield increase on biologics DS downstream

Over 50% reduction of quality deviations

CPV/AQM reporting time reduced by up to

95%

99% time reduction for deviation

investigations

$10’s millions in avoided batch losses

33% reduction of annualized downtime

40

Any Questions?

Please raise your hand and a team member will bring you a microphone.

OR

Scan the QR code to submit your questions using Slido.

We Will be Back After the Break!

Welcome Back to Seeq Summit Mumbai 2026

43

Please Welcome our Next Speaker

FABIO VILLA

CEO SAGAR PRASAD

Sales Director

ERNESTO BOVE

Co-founder and COO

www.e-matica.com marketing@e-matica.com

Trusting operational data: the true prerequisite for reliable

Industrial Advanced Analytics

http://www.e-matica.com/ http://www.e-matica.com/ http://www.e-matica.com/ mailto:marketing@e-matica.com mailto:marketing@e-matica.com mailto:marketing@e-matica.com

Agenda

01 Company Profile

02 How to make trusted operational data?

03 Ematica India

Company Profile

• 25+ years of experience in real-time data management and digital

transformation

• Leveraging real-time data to improve operational processes and the asset

lifecycle

• Over 200 successful projects delivered

• Diverse market experience, including Power, Oil & Gas, Chemicals, Life

Sciences, and Food & Beverage

• Endorsed Partner of AVEVA and Reseller of Canary Labs

• Value-Added Reseller (VAR) for Seeq

Ematica: Experience and Innovation

120+ WORLDWIDE CLIENTS

5 CONTINENTS

50+ PROFESSIONALS & ENGINEERS

100M+ REAL-TIME TAGS MANAGED DAILY

75% LARGE-SCALE CLIENTS

3B€+ IN ANNUAL CLIENT SAVINGS

AI for Industrial

Performance & Sustainability

90%+ REDUCTION IN CLIENT DOWNTIME

Our main clients

Over 75% of our portfolio consists of large-scale, multinational enterprise clients.

• Asahi • Ferrero

Food & Beverage

• Azule Energy • ENI • Shell • SBM Offshore

Oil & Gas

• Arkema • INEOS

• Lamberti • 3M • Solvay

• GSK • Pfizer

• Takeda • Bristol Myers Squibb • Novartis • Olon

• Dalkia • ENGIE

• ENEL • Edison

Power & UtilitiesLife ScienceChemical

The True Prerequisite for reliable Industrial Advanced Analytics

The AI Paradox in Industry

• Heavy investments in sophisticated AI/ML

algorithms

• Over 70% of industrial AI pilots fail to scale

to production

• The problem is the structural layout of raw

data

Why traditional approaches fail

• IT Data vs OT Data: IT is clean by design

(clicks, logs)

• OT Data is born dirty: noisy, scattered, and

disconnected

• Core Law: Garbage In, Garbage Out.

AI cannot magically clean raw mud without

foundational structure

Trusting Operational Data

Why Industrial Data cannot be trusted blindly

2. Lack of Context

Meaningless Raw Numbers

• A timestamp of '85°C' on its own is purely digital noise

• Algorithms need to know the active Batch ID

• Needs correlation with asset state and active product recipe

• Context transforms numbers into actual insights

3. Temporal Integrity

The Cost of Microseconds

• High-frequency machinery requires strict timestamp alignment

• Slight network latency distorts cross-sensor correlation

• Misaligned timestamps lead to phantom anomaly detections

• Analytics require synchronized, uncorrupted edge logging

1. Fragmentation

The Museum of Technology

• Legacy sensors mixed with brand new machinery

• Multi-brand PLCs and locked SCADA systems

• Data exists but sits trapped in isolated operational silos

• No standard communication protocols across lines

The high stakes of untrusted data

• Compromised data = compromised batch A single uncalibrated sensor can render plasma fractions or critical biologic therapeutics completely unusable.

• The Golden Batch failure Real-time predictive models turn into multi-million dollar liabilities if they generate false alerts or completely miss critical operational drift.

• The Compliance nightmare Manual data corrections downstream destroy data lineage, sparking severe regulatory warning letters and factory-floor shutdowns.

Lifescience : zero tolerance for error

DATA IS THE PRODUCT

Strict ALCOA+ Standards FDA CFR 21 Part 11 GAMP 5 Framework

Building the Infrastructure of Trust: AVEVA PI & Canary Labs

AVEVA PI System

Enterprise Data Modeling & Context

▪ Asset Framework (AF): Builds standardized templates (Digital Twins) for repeatable asset types across multi-site production facilities

▪ Event Frames: Automatically layers execution context directly onto time-series data streams (Batch starts, recipe changes, process phases)

▪ Analytics-Ready Out: Feeds completely organized tables directly to corporate machine learning stacks

Canary Labs

High-Performance Edge Logging

▪ Ultra-Fast Sampling: Captures millisecond- level process changes critical to blood fractionation monitoring

▪ Lightweight Edge Architecture: Seamless deployment directly beside production lines with zero data loss risk

▪ Open API Integrations: Publishes clean data via MQTT and modern Web APIs directly into streaming analytics architectures

Build your data trust before your next AI project

Ematica: Bridging the IT/OT divide

Pharma Governance

Enforcing built-in Audit Trails, electronic signatures, validation protocols, and comprehensive compliance with FDA CFR 21 Part 11 and GAMP 5.

Data Shaping

Cleaning, filtering, and structural modeling inside AVEVA PI and Canary so data scientists can deploy algorithms immediately without Excel prep work.

IT/OT Convergence

Connecting raw automation layers (PLCs, SCADA) smoothly with business operations systems (MES, LIMS, ERP) to construct a real single source of truth.

Core Value Proposition: we bring deep domain expertise for different industry sectors

We understand the industrial process and the language of pharma compliance.

Ematica Focus

Ematica India

Speakers’ contact

Ernesto Bove Co-funder CCO Life Science Ematica

ernesto.bove@e- matica.com

Sagar Prasad Sales Director, India Ematica

sagar.prasad@e- matica.com

Fabio Villa CEO Ematica

fabio.villa@e- matica.com

www.e-matica.com

www.e-matica.com

Thank you!

http://www.e-matica.com/

59

Any Questions?

Please raise your hand and a team member will bring you a microphone.

OR

Scan the QR code to submit your questions using Slido.

We Will be Back After the Lunch Break!

Welcome Back to Seeq Summit Mumbai 2026

62

Please Welcome our Next Speaker

SANKAR KANNAN

Head of Technical

63© Indorama Ventures

Optimizing Boiler & Steam Performance Using Seeq

By Sankar Kannan,

Head – Technical Service, TPT Branch 2 Clarity for All. Ownership by All.

64© Indorama Ventures

Who we are!

TPT 2, Thailand

Our Challenge Data

Intelligence Impact

& Benefits

65© Indorama Ventures

Who we are!

TPT 2, Thailand

Our Challenge Data

Intelligence Impact

& Benefits

66© Indorama Ventures

Company Snapshot – Indorama Ventures Limited (IVL)

Lifestyle Fibers

Home Care

Surfactants

Mobility Fibers

PET

Propylene Oxide

Hygiene Fibers and Non- wovens

Personal Care

Surfactants

31 countries

Presence in

*update as of Jul 2026

110 manufacturing locations *update as of Jul 2026

employees

24,500 about Consolidated revenues of

$13.6B (FY2025)

67© Indorama Ventures

Our Global Manufacturing Footprint

*Update as of July 2026

Americas

EMEA

APAC

Brazil Mexico

Uruguay USA

Bulgaria Czech Republic Denmark Egypt France Germany Ghana Italy

Lithuania

Luxembourg

Netherlands Nigeria Poland Russia Slovakia Spain Tanzania Turkey UK

China India Indonesia Israel

Myanmar Philippines Thailand Vietnam

• 110 locations • 31 countries • About ~24,500

colleagues

Indorama Ventures in numbers

Specialty Chemicals

Packaging

68© Indorama Ventures

Our customers help us grow as a market leader

#2Ethoxylation company globally

#1 in PET staple fiber ASEAN

Lifestyle

#2 in BiCo fiber

Hygiene

#2 in airbag

yarn

Mobility

#2 in tire cord

fabric

Mobility

Fibers

#1 Non-ionic surfactants Producer in Americas

#1 mericas’ home care ingredients providerIndovinya

#1 PET producer globally #1 Recycled PET

producer globally

The only shale-integrated player in the West

#1 Crop solutions provider in Americas

Indovida

CPET

Packaging leadership in emerging markets#1 Market

position in ASEAN #2 Market

position in Africa

69© Indorama Ventures

TPT Petrochemicals PCL (Branch 2)

Located in Rayong, Thailand, TPT-2 produces Purified Terephthalic Acid (PTA). Supply raw material for IVL’s PET and Polyester operations and also to external customers worldwide.

CPET Segment 1st Place Winner for Sustainability for CPET

Sankar Kannan Head – Technical service

TPT2 (CPET Segment)

Optimizing Boiler and Steam Performance Using Seeq

Sustainability Project

From Indorama Ventures’ Internal AIA (Seeq) Showcase – Dec. 2025

71© Indorama Ventures

Who we are!

TPT 2, Thailand

Our Challenge Data

Intelligence Impact

& Benefits

72© Indorama Ventures

Boiler & Steam System | Our Biggest Opportunity Driving Energy, GHG & Cost Optimization Across the Steam Value Chain

“True progress lies where business performance and sustainability reinforce one another—driven by purpose, collaborative partnerships, and unwavering commitment.” - YASH LOHIA, Chairman of the ESG Council

77% Energy Consumption

58% Scope 1 GHG Emissions

34% Variable Conversion Costs

Boiler Heat Losses (Fuel gas, radiation, moisture, unaccounted)

Heat Transfer Losses (Fouling, poor approach, low OHTC)

Steam Distribution Losses (Venting, letdown, press. losses)

FUEL

Natural Gas + Biogas

STEAM GENERATION

HP Boilers

STEAM DISTRIBUTION

Headers & Users

PROCESS HEAT TRANSFER

Heat Exchangers

CONDENSATE RECOVERY

Return to Boiler

IVL 2030 TargetsSYSTEM AT A GLANCE Two HP boilers: 85 TPH each at 98 barg

Primary fuel: Natural gas, supplemented by biogas

Existing Heat recovery: Economizer & Air preheater

Primary opportunity: Reduce losses across generation, distribution, & consumption

15% reduction in energy intensity

30% reduction in Scope 1 & 2 GHG intensity

Baseline: 2020

73© Indorama Ventures

Boiler & Steam System | Our Biggest Opportunity Driving Energy, GHG & Cost Optimization Across the Steam Value Chain

“True progress lies where business performance and sustainability reinforce one another—driven by purpose, collaborative partnerships, and unwavering commitment.” - YASH LOHIA, Chairman of the ESG Council

77% Energy Consumption

58% Scope 1 GHG Emissions

34% Variable Conversion Costs

IVL 2030 Targets

Boiler Heat Losses (Fuel gas, radiation, moisture, unaccounted)

Heat Transfer Losses (Fouling, poor approach, low OHTC)

Steam Distribution Losses (Venting, letdown, press. losses)

FUEL

Natural Gas + Biogas

STEAM GENERATION

HP Boilers

STEAM DISTRIBUTION

Headers & Users

PROCESS HEAT TRANSFER

Heat Exchangers

CONDENSATE RECOVERY

Return to Boiler

SYSTEM AT A GLANCE Two HP boilers: 85 TPH each at 98 barg

Primary fuel: Natural gas, supplemented by biogas

Existing Heat recovery: Economizer & Air preheater

Primary opportunity: Reduce losses across generation, distribution, & consumption

15% reduction in energy intensity

30% reduction in Scope 1 & 2 GHG intensity

Baseline: 2020

74© Indorama Ventures

NG Air

BFW

HP steam

Flue gas Process Unit

Boilers

Steam losses

Heat losses

Product

Gap Analysis

KPI visibility without root-cause visibility

Energy losses not consistently quantified

Delayed detection of optimization opportunities

Resource-intensive manual data analysis

HP Steam Boilers Process Unit

STEAM GENERATION STEAM CONSUMPTION

Move beyond "What changed?" to understand “What changed, Why? and What's next?"

• Flue gas temp

• Stack Excess O2

• CO level

• Fuel Composition

• Radiation loss

• Biogas flow

• Biogas HHV

• Blowdown

• BFW temperature

• OHTC / Fouling

• Process duty

• Temperature prediction

• Steam Quality

• Steam Balance

• Venting

• Letdown

• Condensate Recovery

• Unaccounted loss

BOILER EFFICIENCY PROCESS HEAT TRANSFEROPERATING CONDITIONS

Main KPI Reduce NG Specific

Consumption MMBTu/MT of PTA

STEAM DISTRIBUTION

75© Indorama Ventures

Who we are!

TPT 2, Thailand

Our Challenge Data

Intelligence Value

Delivered

76© Indorama Ventures

Define

From Process Data to Energy Intelligence with Seeq

Data Performance Intelligence Value

Connect & Organize • DCS/ Historian data

• Asset groups

• Relevant variables

• Replicate the formulas

Calculate & Monitor • KPI models

• Trends

• Performance metrics

Detect & Diagnose • Heat maps

• Scorecards

• Alerts

Quantify & Optimize • Energy loss

• Opportunity

• Action

• Assess monetary impact ($)

The same framework is applied across boiler, steam and heat-transfer systems.

77© Indorama Ventures

Target: Real time Boiler efficiency through Indirect Method using Seeq

Gap: Natural gas composition analysis not available in site historian

Results: ➢ Real time trend of boiler efficiency. ➢ Drill down individual contributors for efficiency loss.

Steam Generation | Optimizing Boiler Performance Data collection & Organizing Efficiency Calculation – Indirect Method

Seeq supports direct website data

integration via API connections

fx Formula

78© Indorama Ventures

Steam Generation — Biogas Utilization & Heating Value Estimation

Define Target: Predict Biogas Heating Value using Seeq

Limitation: Biogas quality is analyzed only twice a year

Approach:

➢ Predict NG consumption without biogas using relevant variables

➢ Quantify NG displaced by biogas based on actual NG flow.

Outcome: Soft sensor for Biogas Heating Value & Methane% in real time.

AI Agents – Generate calculation Formulas at ease with simple Prompts

User Prompt

Model & Prediction

Utilizing AI Agent to create formula

Biogas Quality Impacting on NG Cons.

To predict NG consumption without biogas

Calculated Biogas Methane%

Biogas flow

Calculated Biogas HHV

NG displaced by biogas

Predicted NG without biogas

NG flow with biogas

79© Indorama Ventures

Steam Consumption — Heat Exchanger Performance & Temp Prediction

Define

Process Flash Steam HP Steam

Heat Exchanger Network OHTC Calculation & Outlet temperature prediction

Goal: Predict Overall Heat Transfer Coefficient (OHTC)

Challenge: Manual calculation / Real time trend not possible

Tools:

➢ tool to configure engineering calculations.

➢ Utilize built-in functions for steam tables & properties

Outcome: Real time trend of Heat exchanger OHTC

Automated alerts for drop in OHTC below desired value

Goal: Predict Heat exchanger outlet temperature

Tools:

➢ feature to predict using relevant variables

Outcome: Soft sensor for exit temperature. Useful for data analysis

fx Formula

Model & Predict

Actual Temperature

Predicted Temperature

Calculated OHTC

Actual close to Prediction

Gap in Actual vs Prediction

80© Indorama Ventures

Steam Consumption — Steam Balance & Loss Identification

HP Steam distribution & Steam Balance Vent / unaccounted loss tracking & Cost Impact

Target: Quantify the cost impact of vent and unaccounted steam losses

Pain point: Delayed action / Difficult to quantify vent & unaccounted loss.

Method:

➢ Calculate vent losses using control valve output

➢ To aggregate the cumulative losses.

Outcome: Quantified energy & cost impact

Drive immediate corrective action to reduce steam losses.

fx Formula

Score Card Metric

Total Vent Quantity

Signal From Condition

Boiler Vent Control Valve opening

Calculated Vent Flow Rate

Total Energy Loss

Cost Impact

Boiler-A vent Boiler-B vent

NG loss via HPS vent

Boiler steam generation Accounted HP steam cons

Unaccounted HP steam loss Vent & Unaccounted loss via HPS

81© Indorama Ventures

Steam Consumption — Process Optimization

Define

Golden Period

Current Period

Traditional Method (Manual Approach)

• Manual comparison of current vs. golden period

• Excel-based calculations for each variable

• Time-consuming and prone to human error

• No automatic linkage to real-time data

• Difficult to repeat or standardize

Seeq Automated (Data-Driven Approach)

• Automated comparison of any two time periods

• Consistent and accurate KPI calculations

• Quickly highlights Impact Variables

• Instant re-analysis just by changing the dates

• Enhanced operational agility and Improved

decision-making

Tree Map to visualize relevant variables

AIA Advantage Vs Traditional Method Benchmarking against Golden Period

Treemap & Score Card Metric

82© Indorama Ventures

Environmental Compliance — Boiler CEMS Monitoring & Alerts

Define

NOx UCL

NOx USL

CO UCL

CO USL Notification

Target: To configure automated alerts for excursions in Boiler stack emission parameters

System : Stack emission data connected live to Industrial authority through Continuous Emission Monitoring System (CEMS) Approach:

➢ Create conditions to automatically generate capsules for potential emission excursions

➢Create automated alerts for capsules generated Outcome: Proactive emission monitoring and compliance with emission norms.

Boiler Stack CEMS Monitoring Automated Alerts for excursions

N O

x C

O

Identify -> Value Search

CO ppm – Online Analyzer

83© Indorama Ventures

Who we are!

TPT 2, Thailand

Our Challenge Data

Intelligence Impact

& Benefits

84© Indorama Ventures

Seeq AIA – Impact & Benefits

The AIA Advantage

• Real-time KPIs. • Early detection of process deviations • Data-driven advanced analytics

• Automated alert for faults • Predictive performance & maintenance • Reduced manual effort in data analysis

• AI / ML powered data analytics • Easily replicated to other entities • AI Agents generate codes with prompts

• Reduce dependency on individuals • Upskill workforce in analytics tools • Faster troubleshooting

• Faster, data-driven decisions • Proactive actions from insights • Insights into measurable impact

Define

20252024

Monthly NG Sp. cons

Yearly Saving in $

Monthly savings in $

Baseline

AIA Adoption

Use Case: Reduction in NG gas Sp. Cons

YTD 2026

1%

100 K$

Cost Saving 10,000 GJ

Energy Saving 580 TCO2e

GHG Reduction

Yearly NG Sp. cons

85© Indorama Ventures

Various Optimization uses of Seeq at TPT2

Daily usage, budget versus actual data, and dashboards provide clear insights for everyone.

Predictive Analytics – Data to Decisions

Risk-focused dashboards → Early anomaly detection → Operating point in curve → Forecasting life of catalyst → Predicting Heat exchanger outlet temp

Automating the Mundane, to Accelerate Action

Organizational Effectiveness, Predictive Analytics, Process Optimization & Sustainability

Process Optimization - No Code ML Models

Optimization accelerated from SPC control to benchmark, Heat Maps & drilldowns, with ML add-ons for advanced analysis.

Plant StatusSustainability Focused Insights

Real-time monitoring of GHG emission (ISO 14064-1) & Energy intensity to drive sustainability goals.

Analytics in Every Hand. Impact at Every Level.

Questions? Sankar Kannan

Head – Technical service TPT2 (CPET Segment)

2026#All_IN Analytics in Every Hand. Impact at Every Level.

Thank You to the Leaders & Team Supporting!

87

Any Questions?

Please raise your hand and a team member will bring you a microphone.

OR

Scan the QR code to submit your questions using Slido.

88

Please Welcome our Next Speaker

CHIRAG SOURABH

Senior Software Engineer

From Machine Learning to Operational Decisions

9 0

The Model is only the beginning A prediction has to make it to the operation before it can create value.

9 1

The last mile of Machine Learning A prediction becomes more valuable when you can understand it in the context of what is happening in the process

9 2

There isn’t one way to get there

9 3

Discover it Start with your operational question. Explore your data. Let the insights guide the path forward.

9 4

From one use case → across the asset fleet

Screenshots of energy transfer case (not real data) (but real seeq)

9 5

Marathon pipeline case

9 6

What if you already have the ML ?

9 7

Bring it Build it there. Use it here.

9 8

Bioreactor • A bioreactor is essentially a controlled tank where

living cells are grown, like a "factory" made of biology instead of machines

• These cells are engineered to produce a specific drug - an antibody, a vaccine protein, etc.

• The tank keeps conditions just right - temperature, oxygen, feed - for 10 to 14 days while the cells do the work

• Sensor probes → this is where live data stream comes from (pH, dissolved oxygen, temperature)

• Feeding pump → controls nutrient delivery, drives the feed rate signal

• Agitation system → keeps cells mixed and oxygenated • System monitor → where all this data is normally just

watched, not predicted - this is the gap ML model fills

9 9

1 0 0

1 0 1

Build it Seeq Data Lab gives you the flexibility to build, automate and extend – your way.

1 0 2

And you’re not building alone.

1 0 3

Different models. Different customers. One pattern.

One thing in common: The prediction is put to work.

1 0 4

The goal isn’t more models. It’s better decisions.

The success of an ML project isn't measured by how sophisticated the model is.

It's measured by what changed because that prediction was available at the right time.

1 0 5

CHIRAGASOURABH

Product Owner | Senior Software Engineer

Seeq Corporation

Thank you

10 7

Book a Meeting With Seeq

Thank you for joining us at the Seeq Summit

Mumbai.

If you’d like to continue the conversation, discuss

a specific topic, or learn more about how Seeq could support your organisation, please scan the QR Code and complete the form.

A member of the Seeq team will contact you

within the next few business days to follow up and arrange a meeting.

Thank You for Joining Seeq Summit Mumbai 2026

Intro Slides Slide 1: Welcome to Seeq Summit Mumbai 2026 Slide 2: Agenda Slide 3 Slide 4 Slide 5: Points to Note During the Day Slide 6: Ask Your Questions Using Slido Slide 7

Dustin Slide 8: Please Welcome our Next Speaker Slide 9: Human Expertise, Amplified Slide 10: The pressure on operations has never been greater. Slide 11: “We’re drowning in data, but starved for information.” Slide 12 Slide 13: Built By Industry. Focused On Experts. Slide 14 Slide 15: Enterprise Monitoring Slide 16 Slide 17: AI Activates Decision Capital Slide 18 Slide 19 Slide 20 Slide 21: Human Expertise Is The Advantage. Amplify it. Slide 22: Any Questions?

Paolo Slide 23: Please Welcome our Next Speaker Slide 24: From Data to Decision: Right First Time and Higher Efficiency with Seeq in Pharma Slide 25 Slide 26: The Problem: Raw Data -> Quality Decisions Slide 27 Slide 28: Seeq for Pharma supporting compliant decisions Slide 29: What decisions? Slide 30: From Blind Spots to Actionable Insight Slide 31: One GMP decision depends on more than one data stream Slide 32: First, make the batch understandable Slide 33 Slide 34: Step 2 – Define your Control Strategy Slide 35 Slide 36: Step 3 – Scale and Monitor by exception Slide 37 Slide 38: This is now Slide 39 Slide 40: Any Questions?

Break Slide 41: We Will be Back After the Break! Slide 42: Welcome Back to Seeq Summit Mumbai 2026

Ematica Slide 43: Please Welcome our Next Speaker Slide 44 Slide 45 Slide 46 Slide 47 Slide 48 Slide 49 Slide 50 Slide 51 Slide 52 Slide 53 Slide 54 Slide 55 Slide 56 Slide 57 Slide 58 Slide 59: Any Questions?

Lunch Filler Slide 60: We Will be Back After the Lunch Break! Slide 61: Welcome Back to Seeq Summit Mumbai 2026

Indorama Slide 62: Please Welcome our Next Speaker Slide 63: Optimizing Boiler & Steam Performance Using Seeq Slide 64 Slide 65 Slide 66: Company Snapshot – Indorama Ventures Limited (IVL) Slide 67: Our Global Manufacturing Footprint Slide 68: Our customers help us grow​ as a market leader Slide 69: TPT Petrochemicals PCL (Branch 2) Slide 70 Slide 71 Slide 72: Boiler & Steam System | Our Biggest Opportunity Slide 73: Boiler & Steam System | Our Biggest Opportunity Slide 74: Gap Analysis Slide 75 Slide 76: From Process Data to Energy Intelligence with Seeq Slide 77: Steam Generation | Optimizing Boiler Performance Slide 78: Steam Generation — Biogas Utilization & Heating Value Estimation Slide 79: Steam Consumption — Heat Exchanger Performance & Temp Prediction Slide 80: Steam Consumption — Steam Balance & Loss Identification Slide 81: Steam Consumption — Process Optimization Slide 82: Environmental Compliance — Boiler CEMS Monitoring & Alerts Slide 83 Slide 84: Seeq AIA – Impact & Benefits Slide 85: Various Optimization uses of Seeq at TPT2 Slide 86 Slide 87: Any Questions?

Untitled Section Slide 88: Please Welcome our Next Speaker Slide 89: From Machine Learning to Operational Decisions​ Slide 90: The Model is only the beginning Slide 91: The last mile of Machine Learning Slide 92: There isn’t one way to get there Slide 93: Discover it Slide 94: From one use case → across the asset fleet Slide 95 Slide 96: What if you already have the ML ? Slide 97: Bring it Slide 98: Bioreactor Slide 99 Slide 100 Slide 101: Build it Slide 102: And you’re not building alone. Slide 103: Different models. Different customers. One pattern. Slide 104: The goal isn’t more models. It’s better decisions. Slide 105 Slide 106 Slide 107: Book a Meeting With Seeq

End Slide Slide 108: Thank You for Joining Seeq  Summit Mumbai 2026