Approach Document

A Progression Framework
for the Clariant GCC

Perspectives on evolving the Mumbai Shared Services Centre into a next-generation Global Capability Centre

3
Progression Dimensions
4
Core Functions
1
AI Platform Already Built
Understanding the Client

Clariant at a Glance

A Swiss multinational specialty chemicals company headquartered in Muttenz, near Basel. Formed in 1995 as a spin-off from Sandoz, operating across Care Chemicals, Catalysts, and Adsorbents & Additives.

๐Ÿข
CHF 4.4B
Revenue (2023)
๐ŸŒ
10,465
Employees globally
๐Ÿ“
36
Countries
โš—๏ธ
3
Business Units

Global Business Services

Three SSCs in Mumbai (India), Lodz (Poland), and Dalian (China). Covers Order-to-Cash, Source-to-Pay, and Hire-to-Retire spanning AP, AR, procurement, reporting, master data, logistics, and credit risk management.

Clarita: The GenAI Platform

Built on AWS Bedrock and Anthropic Claude. Launched early 2023 and rolled out to 1,000+ employees. A modular platform covering R&D, Sales, and corporate reporting. Ranked #1 in VAA Chemistry Managers survey for AI adoption among 21 companies.

Current State

The Mumbai SSC Today

Serving English-speaking countries across Asia Pacific and North America. The centre handles standardised, transaction-based back-office processes across four primary functions.

๐Ÿ“Š

Finance

Accounts Payable & Receivable
Invoice Processing
Credit Risk Management
Financial Reporting
๐Ÿ“‹

Procurement

PR to PO Conversion
Policy Compliance Matching
Three-Way Invoice Matching
Master Data Management
๐Ÿ‘ฅ

HR

Employee Lifecycle Admin
Leave & Attendance
Payroll Support
Compliance & L&D Tracking
๐Ÿ–ฅ๏ธ

Shared IT

IT Help Desk
User Self-Service
Infrastructure Documentation
Application Support
๐Ÿค–

RPA Already in Progress: UiPath deployed for invoice processing (~50% automation in pilot region) and logistics, saving 100+ hours/month and eliminating 40,000 printouts per quarter.

MUMBAI SSC COVERAGEFinanceAP, AR, Invoicing, RiskProcurementPO, Matching, MDMHRLifecycle, Leave, PayrollShared ITHelp Desk, Self-ServiceRPA Layer: UiPath (~50% invoice automation, 100+ hrs/mo saved)MUMBAI SSC COVERAGEFinanceAP, AR, Invoicing, RiskProcurementPO, Matching, MDMHRLifecycle, Leave, PayrollShared ITHelp Desk, Self-ServiceRPA Layer: UiPath~50% invoice automation, 100+ hrs/mo saved
The Framework

Three Dimensions of Progression

This is not about growth in headcount. It is about progression in capability, intelligence, and strategic relevance. Click a dimension below to explore it in detail.

Where we focus today

We recognise how much ground Clariant has already covered. Our understanding is that levelling up the centre (01) and scaling automation (02) are well thought through, and either under way or already delivered, with RPA at an advanced stage. So rather than revisit familiar territory, today's conversation centres on Dimension 03: the agentic transformation.

01
Level Up
Value-Added Functions
โœ“In motion
Click to explore โ†“
02
Automate Further
Intelligent Automation
โœ“Well advanced
Click to explore โ†“
03
AI Transformation
The UnIt Model
Today's focus
Click to explore โ†“
โœ“ In motion at Clariant

Our understanding is that this progression is already well thought through, with much of it under way or achieved. It is included here to keep the full framework in view.

Dimension 01: Level Up

Each function handles standardised, rule-based work today. The opportunity is to absorb higher-intelligence activities that currently reside at headquarters.

The expertise already exists within the team. As seniority grows and domain familiarity deepens, these transitions become natural rather than forced.

PROGRESSION JOURNEYTransactionalRule-based, repetitiveAnalyticalInsights, benchmarksStrategicAdvisory, planningCurrentNear-termTargetPROGRESSION JOURNEYTransactionalRule-based, repetitiveCURRENTAnalyticalInsights, benchmarksNEAR-TERMStrategicAdvisory, planningTARGET
โœ“ Well advanced at Clariant

With UiPath firmly established and RPA at an advanced stage, this dimension is largely in place. It is included here to keep the full framework in view.

Dimension 02: Automate Further

Clariant's GBS has already invested in UiPath and achieved meaningful results. The next step is scaling coverage and layering intelligence.

~50%
Invoice processing automated
100+
Hours saved per month
40K
Printouts eliminated/quarter
โš™๏ธ
Current
Basic RPA
Rule-based bots handling structured, repetitive tasks with human fallback for edge cases.
๐Ÿง 
Next
Intelligent Automation
Adding OCR, NLP, and machine learning for semi-structured data. Smarter exception handling.
๐Ÿš€
Target
Cognitive Automation
Process-aware automation with decision intelligence. Integration with Clarita for predictive actions.
Today's focus

This is the heart of today's discussion: how digital agents, working alongside your people, could reshape the way work gets done across the GBS.

Dimension 03: AI-Infused Transformation

This is the dimension that fundamentally changes the operating model. It reimagines how work gets done at a structural level.

The Core Idea: The UnIt Model

Instead of measuring capacity in FTEs alone, think in terms of units: U (the human) + It (the agent), working together. Each person manages and orchestrates multiple AI agents, expanding effective capacity by 3x to 10x.

U
Human
+
It
Agent
=
UnIt: the new capacity measure
Today
"We need to go from 100 to 150 FTEs"
โ†’
Tomorrow
"Each of our 100 people manages 3-5 agents"
THE UnIt OPERATING MODEL๐Ÿ‘คHR AnalystOrchestrator๐Ÿ“…Leave Mgmtโœ…Compliance๐Ÿ“šL&D Tracking๐Ÿš€OnboardingEach agent inherits role-based access from its human orchestratorTHE UnIt OPERATING MODEL๐Ÿ‘คHR AnalystOrchestrator๐Ÿ“…Leave Mgmt๐Ÿš€Onboardingโœ…Compliance๐Ÿ“šL&D TrackingEach agent inherits role-based accessfrom its human orchestrator

Starting Smart: Initial deployment focuses on processes where an agent error has no direct financial impact. Leave management, compliance tracking, and L&D administration are ideal starting points. Payroll and compensation come later once trust and reliability are established.

The Platform

Clarita: Already Built, Already Working

Clariant already possesses one of the most capable enterprise GenAI platforms in the specialty chemicals industry. Clarita is not a pilot; it is production-grade with proven adoption.

โšก

Foundation

Built on AWS Bedrock and Anthropic Claude. Modular architecture allows new GenAI projects to launch in under 4 hours.

๐Ÿ”ฌ

R&D Acceleration

Synthesises scientific papers, patents, and toxicology data. Helps researchers identify insights across Clariant's vast knowledge base.

๐Ÿ’ผ

Sales Intelligence

Clarita Sales Assistant provides customer insights, technical guidance, and business development support for informed conversations.

๐Ÿญ

Operations

Reduced steam consumption at a German plant and optimised energy usage at an Indonesian mining site by analysing operational data.

๐Ÿ“

Corporate Reporting

Used for integrated annual report creation, crafting product stories, and simplifying complex sustainability narratives.

๐Ÿ”’

Data Security

Operates only within Clariant's network. Every API on a separate AWS account, datasets fully isolated. Does not learn from user interactions.

Ranked #1 of 21 companies (VAA Survey)1,000+ active users globally"Coffee with Clarita" sessions

The Implication

The platform infrastructure for AI-infused transformation already exists within Clariant. Clarita's modular design means the capabilities for agent orchestration can be configured, not built from scratch. This is a significant head start.

Controls

Governance, Security & Digital Identity

Agents operate under the same governance framework as humans. No exceptions, no shadow access, no untracked actions.

๐Ÿชช

Agent Identity

Each agent receives its own unique ID within the enterprise identity system, similar to how BNY Mellon's digital employees are structured. Agents are traceable and individually accountable.

๐Ÿ”

Inherited Access Control

An agent's system access mirrors the permissions of the human who creates it. A mid-level manager's agents inherit that manager's access scope, nothing more.

๐Ÿ“‘

Full Audit Trail

Every agent action generates the same logs and audit records that a human action would. Who accessed what, when, and what changed, all captured automatically.

๐Ÿ’ฌ

Agent-Human Interaction

Agents can raise queries when they encounter ambiguity. The human orchestrator reviews, responds, and redirects, maintaining oversight at every step.

Engagement Model

Where We Come In

A clear delineation of what we bring and what stays within Clariant's domain.

Centred on Dimension 03

With levelling up and automation already in motion at Clariant, our contribution is shaped around the agentic transformation: the area we are here to explore with you today.

YASH Brings

โœ“
Design Thinking framework to identify and prioritise the processes that are ready for agentic AI, scored on volume, decision type, data readiness and risk
โœ“
Deep agentic AI expertise across the lifecycle: from POV, to a sandbox production run in shadow mode, to production deployment with continuous monitoring (AgentOps)
โœ“
Unified Workforce operating blueprint that plans humans and digital agents as one capacity model, with defined roles and career paths
โœ“
Process mapping and change management methodology that redesigns work at task level and drives adoption across affected teams
โœ“
Governance framework for digital agent deployment: autonomy tiers, guardrails, audit trails and compliance controls
โœ“
Benchmarks and reusable patterns from GCC and GBS engagements to accelerate time-to-value

Clariant Owns

โœ“
Clarita platform configuration and development
โœ“
Internal technology implementation and deployment
โœ“
Data architecture and security infrastructure
โœ“
Business process execution and operations
โœ“
Internal stakeholder alignment and buy-in
โœ“
Ongoing platform evolution and maintenance

We are not claiming to touch or modify the Clarita platform. We bring the concepts, the frameworks, and the enablement methodology.

Today's Discussion

From Transaction Centre to Intelligence and Orchestration Centre

Process Intelligence. Smarter Execution. Greater Impact.

PeopleProcessesTechnologyworking together for a more sustainable world

Not a bigger GBS.
A more valuable one.

The GBS Progression

From execution to outcome ownership
โš™๏ธ
STEP 1
Execute
  • Process transactions
  • Meet SLAs
  • Follow rules
๐Ÿ“Š
STEP 2
Understand
  • See process variants
  • Identify bottlenecks
  • Quantify value leakage
๐Ÿ’ก
STEP 3
Recommend
  • Diagnose root causes
  • Propose improvements
  • Provide evidence
๐Ÿ”—
STEP 4
Orchestrate
  • Coordinate people, systems, bots and AI
  • Manage exceptions
  • Track to closure
๐ŸŽฏ
STEP 5
Own
  • Take outcome accountability
  • Continuously improve
  • Be a strategic partner
๐Ÿงญ
The Goal

A GBS that understands, improves and orchestrates global processes using the right mix of people, applications, RPA and AI.

The Foundation

Clariant GBS Today

A strong foundation to build on.

๐Ÿ’ฐ

Finance

  • AP & AR
  • Invoice processing
  • Credit risk management
  • Financial reporting
  • Master data
๐Ÿ›’

Procurement

  • PR to PO conversion
  • Policy compliance
  • Three-way matching
  • Supplier master data
๐Ÿ‘ฅ

HR

  • Employee lifecycle
  • Leave & attendance
  • Payroll support
  • Compliance & L&D
๐Ÿ’ป

Shared IT

  • Help desk
  • User self-service
  • Infrastructure docs
  • Application support
๐Ÿค–
RPA in progress: UiPath (~50% invoice automation) and logistics automation.
100+
hours saved
per month
40,000
printouts eliminated
per quarter
Already in Place

Clarita: A Strategic Advantage

An enterprise GenAI platform already in place.

โ˜๏ธ
Built onAWS Bedrock and Claude
๐Ÿ‘ฅ
1,000+
active users globally
๐Ÿ†
Ranked #1
in VAA survey (21 companies)

Used across the enterprise

๐Ÿงช
R&D Acceleration
๐Ÿ“ˆ
Sales Intelligence
๐Ÿญ
Operational Optimisation
๐Ÿ“„
Corporate Reporting
Secure. Modular. Production-grade. Ready to extend to GBS workflows.
Where the Value Lies

The Opportunity

Not more headcount. Greater capability, intelligence and relevance.

๐Ÿ“ˆ
01
Move up the value chain
From transactions to analysis and advisory.
โš™๏ธ
02
Use the best-fit execution mechanism
For each activity.
See the decision tree โ†’
๐Ÿ‘ฅ
03
Redesign work as UnIts
Human + It.
See the UnIt model โ†’
๐ŸŽฏ
04
Create measurable business outcomes
Cost, cash, service, control and continuity.
Decision Framework

Best-Fit Execution Decision Tree

Start with the business problem. Choose the simplest, most sustainable solution.

๐Ÿงญ

Ask the questions in order and stop at the first good fit: the earlier the answer, the simpler and more sustainable the solution. Whatever the route, step 5 decides what stays with a person.

0
Should this activity exist?
Add value?Policy or control?Can it be eliminated or simplified?
Eliminate / Simplify / Standardise
1
Can the core application handle it?
SAP capability?Configuration issue?Master data?S/4HANA roadmap?
Strengthen Enterprise Application
2
Can workflow or integration solve it?
System handoffs?APIs available?Routing / approvals?Rekeying / reconciliation?
Use Integration or Workflow
3
Is it deterministic and repeatable?
Structured inputs?Clear rules?Stable volume?Limited exceptions?
Use RPA / Deterministic Automation
4
Does it require language, context or prediction?
Unstructured inputs?Variable exceptions?Need for interpretation?Knowledge retrieval?
Use AI / Agentic Augmentation
5
What must remain with a person?
Financial impact?Compliance or risk?Irreversible decision?Ethical judgement?Accountability?
Human-led UnIt

Try it on a real process

Pick one process and answer as you go. It takes under a minute.

Five Lenses

Key Decision Parameters

Evaluate each opportunity across five lenses.

๐Ÿ“Š
Business Value
CostCashCycle timeServiceControlContinuity
โš™๏ธ
Process Characteristics
VolumeVariabilityRule clarityException rateStandardisationStability
๐Ÿงฉ
Technology Fit
SAP capabilityAPIsUiPath fitClarita fitData readinessS/4 alignment
๐Ÿ›ก๏ธ
Risk & Governance
Data sensitivityFinancial consequenceExplainabilityAuditabilityHuman accountability
โณ
Long-Term Viability
Total costMaintainabilityChange frequencyReusabilityPlatform longevity
Operating Model

The UnIt Model

Human + It = Greater Outcomes.

U ยท Human Intelligence

  • Judgement
  • Context
  • Accountability
  • Exception handling
  • Relationships
UHumanIntelligenceItDigitalExecutionUnItOutcome-focused

It ยท Digital Execution

  • Applications (SAP)
  • Workflow / Integration
  • RPA
  • AI / Agents
  • Analytics & Monitoring
UnIt

Outcome-focused operating capability

Hypothesis

Potential Starting Point

A hypothesis to test together, not a conclusion.

Hypothesis

Source-to-Pay Exception Intelligence

๐ŸŽฏ

Why it's compelling

  • Direct link to working capital and cost management
  • Material exception volume and rework
  • Touches SAP, workflow, RPA, AI and human judgement
  • Builds on existing RPA success
  • Replicable across regions and business units
๐Ÿ”

Candidate journeys to explore

  • PR completeness and policy routing
  • PR to PO exceptions
  • Catalogue vs. non-catalogue buying
  • Supplier master data exceptions
  • Three-way match and invoice blocks
  • Repeated vendor enquiries
๐Ÿ“‹

To be validated in discussion

  • Is S2P already under active redesign?
  • Where is the biggest friction?
  • What has been addressed through UiPath?
  • What data is available for a baseline?
  • Who owns the process?
  • Would another tower be a better first step?
Proposed First Step

GBS Process Intelligence and UnIt Discovery Sprint

From a shortlist of workflows to one bounded proof, with success measures and a scale path agreed up front.

1

Select 2-3 candidate workflows

2

Evaluate using the value and readiness framework

3

Choose one process for a bounded discovery and proof

4

Define baseline, target workflow, execution design and UnIt roles

5

Agree on success measures and scale path

From This Discussion

Desired Outcome

Where we would like to land by the end of today's conversation.

๐Ÿค
Agreement to
  1. 1Nominate 2-3 candidate processes
  2. 2Bring the right stakeholders
  3. 3Apply the decision framework together
  4. 4Select one process for a focused discovery sprint
  5. 5Define next steps and timeline

Together, we can turn process excellence into a competitive advantage.

Clariant | YASH | A More Intelligent Tomorrow

The Art of the Possible

Clariant's Mumbai centre has a strong foundation, a progressive leadership mindset, and an AI platform that most GCCs would envy. The ingredients for transformation are already in place.

01
Level Up
Higher-value functions
02
Automate
Intelligent automation
03
UnIt
Human + Agent workforce
YASH Technologies | GCC Advisory | Confidential